WO2006002688A2 - Incompatibility processing - Google Patents

Description

Specification

INCOMPATIBILITY PROCESSING

The present invention relates to the field of transportation planning and optimization (TPO). In general, the goal of transportation planning and optimization is to find a feasible transportation plan with minimum costs. A transportation plan assigns transportation or¬ ders to vehicles and determines the scheduling of the involved logistic activities. Such a plan is called feasible, if it satisfies the desired technical constraints, like e.g. the loading capacities of vehicles. There is a large variety of different constraints which may be de¬ noted as compatibilities or incompatibilities.

For example, certain product groups may not be shipped together on the same vehicle, or¬ ders of specific incoterms, i.e., specific contractual trade regulations, are not to be com¬ bined into same shipments, locations may not have special equipment to unload specific products, etc.

Compatibility can not only be (in)compatibility in itself, but also, for example, a require¬ ment for special transportation conditions like vehicle, or the definition of a specific indi¬ rect shipment mode in multimodal transportation. The last possibility makes this especially important in transportation scenarios involving multi-cross-docking. These constraints should be taken into account efficiently during TPO.

As mentioned above, the main aim of TPO is to create an optimal transportation plan, which contains in particular route, scheduling and capacity information. The main objects which are involved in TPO and which are important from compatibility point of view are: order: represents transportation order in form of freight unit, which contains all plan- ning-relevant information, derived from original order document (like sales order, de¬ livery, stock transfer order, etc.) information. Order number uniquely identifies corre¬ sponding order in planning system and can be used to get data for this order. - vehicle: represents capacity, availability and other information which corresponds to physical vehicles performing transportation (trucks, rail, etc.)

- vehicle: represents the class of vehicles with similar or identical properties (from plan¬ ning perspective). cross-docking location : represents possible location (warehouse, plant, carrier, distri- bution center, etc.), where goods can be reloaded. Cross-docking location number uniquely identifies corresponding location in planning system.

The result of TPO are shipments, which contain information on assignments of orders to vehicles, corresponding activities (like goods issue, receipt, transport, empty leg and so on) and other planning information like stages, carriers, deadlines, etc.

In TPO systems of the art, there is no general concept of modeling compatibilities. Only very specific incompatibility requirements can be modeled in known transportation plan¬ ning/vehicle scheduling (TP/VS) systems: Vehicle - Location, Transportation Group - Transportation Group, Vehicle - Transportation Group. These compatibilities are defined explicitly on corresponding fields. They are maintained in separate transactions, also proc¬ essing of them is done in different areas. Optimizer interface contains three input tables, which directly correspond to these incompatibility structures. Transportation group is specified as field in order data input data.

Further, there is an impossibility to add new compatibility requirements without changing optimization engine and extending the interface. Further, it is not possible to use order characteristic other than transportation group, for compatibility definition. Moreover, it is not possible to restrict validity of compatibilities to orders which satisfy certain criteria. Yet more, there is a separate maintenance of compatibilities of different types.

One object of the present invention is to overcome the restrictions as described above. This object is achieved by the methods and apparatuses according to the claims.

According to the embodiment of the invention, there is provided a method of modeling incompatibilities in a transportation planning and optimization system, wherein a transpor¬ tation plan is determined for

a plurality of orders, each order being defined by a predefined number of character¬ istics, one of the characteristics specifying freight units to be transported from a predetermined place of departure to a predetermined place of destination, each freight unit representing a predefined good, a plurality of vehicles for transporting freight units, and a plurality of cross-docking locations for loading/unloading freight units onto/from vehicles,

the transportation plan defining shipments, each shipment specifying the vehicles, and cross-docking locations needed for fulfilling the plurality of orders, whereby a number of freight units, vehicles, and cross-docking locations are incompatible to each other such that they are not allowed within the same shipment;

each of the freight units, vehicles, and cross-docking locations being specified by attributes which are comprised in a field catalogue;

the modeling method comprising the following steps:

providing a data structure for the field catalogue, the field catalogue providing data field attributes, each of the attributes representing one of

- freight units,

- vehicles, and - cross-docking locations, respectively; providing a data structure for a plurality of compatibility types, the data structure provid¬ ing for each compatibility type data fields for:

- a first entry for linking the compatibility type structure with a first attribute of the field catalogue; a second entry for linking the compatibility type structure with a second attribute of the field catalogue;

providing a number of different compatibility types, each compatibility type being de- scribed by values of the data fields as defined in the provided compatibility type data structure;

whereby a number of compatibility rules is provided, each compatibility rule specifying a combination of two values which refer to first and second attributes of a defined compati- bility type, and being indicative whether or not they are compatible,

the compatibility rules being applied upon determining compatible combinations of freight units, vehicles, and cross-docking locations for the shipments of the transportation plan.

In a further aspect of the invention, the compatibility type data structure further provides a data field for an ID for identifying the respective compatibility type.

The compatibility type ID may be a numeric value.

The compatibility types may comprise at least one of

- transportation group, and transportation group;

- vehicle, and transportation group;

- incoterm, and incoterm; - cross-docking location, and transportation group;

- delivery priority, and cross-docking location;

- delivery priority, and vehicle; - shipping conditions, and cross-docking location;

- shipping conditions, and vehicle;

- weight/volume group, and vehicle;

- weight/volume group, and cross-docking location; - transportation zone, and vehicle; and

- cross-docking location, and vehicle;

A compatibility rule may further comprise a condition ID for restricting the applicability of the compatibility rule onto predefined orders,

According to a further embodiment of the invention there is provided a method of process¬ ing incompatibilities for a transportation plan for

a plurality of orders, each order being defined by a predefined number of character- istics, one of the characteristics specifying freight units to be transported from a predetermined place of departure to a predetermined place of destination, each freight unit representing a predefined good, a plurality of vehicles for transporting freight units, and a plurality of cross-docking locations for loading/unloading freight units onto/from vehicles,

the transportation plan defining shipments, each shipment specifying the vehicles, and cross-docking locations needed for fulfilling the plurality of orders, whereby a number of freight units, vehicles, and cross-docMng locations are incompatible to each other such that they may not be allowed within the same shipment;

each of the freight units, vehicles, and cross-docking locations being specified by attributes which are comprised in a field catalogue;

the method comprising the following steps:

Receiving data which is descriptive of the orders, data which is descriptive of the available vehicles, and data which is descriptive of the available cross-docking locations;

reading a set of compatibility types, and a set of compatibility rules with respect to the plu¬ rality of orders, vehicles, and cross-docking locations, each compatibility rule specifying a combination of two values which refer to first and second attributes of a defined compati¬ bility type, and being indicative whether or not they are compatible;

associating a value with each order, the value being descriptive of the characteristics of the order, such that orders having the same characteristics have the same value associated therewith;

replacing the data descriptive of the characteristics by the value associated therewith; and

outputting the order data resulting from the preceding step.

According to a further aspect of the embodiment, the following steps are performed prior to the outputting step:

classifying orders having the same value associated therewith into a group;

replacing the groups which have identical incompatibility relations to all other groups by a class; and

creating a data representation of the orders, whereby the orders are grouped into classes of orders having the same incompatibilities to the other classes.

According to a further aspect the groups of orders are represented by an object-oriented model, the model comprising as one object the group of orders, and as second object an incompatibility matrix, the matrix representing the incompatibilities between pairs of groups; as third object a data structure representing incompatibilities between vehicle and orders, and between orders and cross-docking locations; and as fourth object incompatibili¬ ties between cross-docking locations and vehicles.

According to a yet further aspect of the embodiment, each order object comprises a data field which specifies the group; the incompatibilities matrix object comprises for each pair of order groups an information as to whether they are compatible or not; each order object comprises a set of legs referring to relevant compatible cross-docking locations; and leg objects which comprise a set of compatible vehicles.

Thus, according to the invention, a generic model is proposed that allows maintenance and processing of compatibility constraints between arbitrary planning object attributes. This model includes:

a new concept of compatibility type, which allows flexible definition of compatibility constraints between arbitrary object attributes; a new algorithm, which performs conversion of such compatibility constraints into four incompatibility classes relevant for transportation, taking into account possible validity conditions; - new optimization algorithms that reduce this generic model into a very compact model, enabling a transportation optimizer to focus on the core of the generic com¬ patibility model.

The present invention allows very generic modeling in a straightforward manner, without loosing efficiency in the optimization and planning. The main advantages of the present invention are:

High modeling flexibility of using of any order attribute as possible compatibility defini¬ tion in combination with vehicle and cross-docking location. This allows definition and handling of arbitrary compatibilities between different order attributes/characteristics, lo- cations, and vehicles. Furthermore, in optimization the general model can be reduced efficiently into its core be¬ fore actual optimization. Thus, optimization does not loose efficiency, but it can solve transportation scenarios of a broader scope, viewed from the perspective of modeling compatibilities.

Stated otherwise, on the one hand, the user describing a transportation optimization prob¬ lem is very flexible because he is not restricted to the compact modeling used by the opti¬ mizer - instead, he can use the very generic modeling approach. On the other hand, the developer of the optimizer has no need to care about numerous different compatibility con- straints. The developer can focus on implementing the four atomic compatibility con¬ straints.

As the invention uses the same compatibility engine, it is achieved that while optimization and manual planning, the same constraints are respected. In manual planning nevertheless they could be disregarded by planners, thus, allowing a common handling of compatibil¬ ities in both optimization and manual planning.

It is not necessary to do any code modifications to handle customer-defined fields included in freight unit, they are processed automatically by standard coding, thus leading to an eas- ier way of handling with customer extensions.

Advantageous implementations can include one or more of the following features.

(Final version of dependent claims to be inserted here)

In particular, the invention comprises also computer systems for performing the inventive methods.

Furthermore, the invention comprises computer-readable storage media comprising pro¬ gram code for performing the inventive methods, when loaded into a computer system.

The present invention is further described with reference to the drawings, wherein Fig. 1 illustrates the definition of compatibility types;

Fig. 2 illustrates the definition of compatibility rules;

Fig. 3 illustrates the processing of compatibilities in TPO; Fig.4 illustrates an overview of the different phases of incompatibility reduction processing according to the invention;

Fig. 5 illustrates the first phase of the reduction processing;

Fig. 6 illustrates the second phase of the reduction processing;

Fig. 7 illustrates the third phase of the reduction processing; Fig. 8 illustrates a general diagram of compatibility processing;

Fig. 9 illustrates a first subprocess of the compatibility processing;

Fig. 10 illustrates the subprocess of preparing incompatible combinations;

Fig. 11 illustrates the subprocess of processing incompatibilities with conditions;

Fig. 12 illustrates the subprocess of classifying incompatible combinations to 4 classes;

Figs. 13 to 16 illustrate the feasible path finding in a hub network model in the different phases of processing;

Figs. 17 to 19 illustrate screen shots of compatibility transaction.

Compatibility type definition is based on field catalogue functionality (see Fig. 1). Field catalogue 100 contains declarations (in form of character identifiers (ID) plus additional optional information like description, check table name etc.) of various attributes of vari¬ ous object types, like order attribute/characteristic, which can be used in particular for compatibility type 200 and condition definitions.

Mainly, compatibility type consists of compatibility type numeric identifier 230 with de¬ scription and two field identifiers from field catalogue 100, which represent (in)compatible combination of object attributes 210, 220. The linkage between field ID 230 and corre¬ sponding object attribute value is provided by field catalogue functionality. All possible compatibility types 200 valid for particular business scenario are identified and defined at customizing step. In a specific example, the following compatibility types may be required: 01- transportation group - transportation group;

02- incoterm - incoterm;

03- vehicle - transportation group; 04- cross-docking location - transportation group.

A compatibility rule 300 (also denoted as "(incompatibility") represents the actual combi¬ nation of two values which refer to objects defined by compatibility type 200 and repre¬ sents the objects which may be considered as (incompatible in TPO (see Fig. 2). Specific compatibility rules 300 are based on compatibility types. In particular, possible first and second attribute values in compatibility rule should be consistent with first and second ob¬ ject types in compatibility type 200. For example, if first attribute of compatibility type 200 is cross-docking location 30, then first attribute values in corresponding rules 300 can only contain values which identify location. Nevertheless patterns can also be used to accelerate maintenance.

On planning/customizing step necessary compatibility rules 300 (incompatibilities) are maintained.

The linkage between attribute in compatibility type 300 and order attribute is realized in the way that attribute ID in field catalogue and therefore in compatibility type coincides with corresponding component in the structure of freight unit 10, or, represents vehicle 20, or, cross-docking location 30. Below, vehicles 10 are sometimes also denoted as means of transport. As freight unit structure 10 already contains most of shipping, product and other information of order, only one read of freight unit entry with compatibility attribute ID would be in most cases enough to get information, whether particular freight units contain (in)compatible attributes, and therefore should be handled correspondingly. In this way, not only any freight unit standard field can be used as possible (incompatibility attribute, but also additional fields in customer include can be also used and processed automatically. If this is not the case (freight unit does not contain attribute), it is still possible to enter value for attribute in user exit provided by TPVS functionality. Technically, one function module is used as compatibility engine for retrieving and proc¬ essing them. It gets orders (in form of freight units), vehicles 20 and cross-docking loca¬ tions 30 as input objects (see Fig. 3). It provides output if form of incompatible combina¬ tions of attributes classified to four classes:

- vehicle - order attribute order attribute - order attribute

- cross-docking location - order attribute

- vehicle - cross-docking location.

These classes represent all possible combinations of incompatible objects, which are rele¬ vant for transportation.

This information can be later used by a transportation optimizer system, or in manual plan- ning in similar way.

In addition to the incompatible attribute combinations, the optimizer also receives the at¬ tributes of all orders. One value is given for each order and possible attribute. All these attributes are described in a separate table.

The transportation optimizer has to determine feasible transportation plans, i.e. plans that satisfy all incompatibilities declared in above listed four classes. Advantageously, the ap¬ proach according to the present invention provides rich modeling capabilities since many constraints involving several different order attributes can be declared by only these four incompatibility classes.

The question is how to solve this optimization problem efficiently, taking into account all different constraints and order attributes. The straightforward approach would be to extend the optimizer's internal search routines by introducing the order attributes and considering them explicitly during the search process. This approach has the drawback that the search logic of an existing optimizer system has to be changed. It is based on local search, which uses lots of different atomic moves. These moves perform small changes in an existing transportation plan, and the selection pressure of local search iteratively applies moves that reduce the costs of the current transportation plan. Thus, in order to consider order attrib¬ utes inside this search process, they must be integrated into all moves. All this would cause a full revision of the internal search process. Furthermore, the search process would not exploit similarities or symmetries in the structure of the order attributes. Thus, search would solve a very general model, without exploiting special structural properties in the incompatibility structures. In general, this approach of integrating the attributes directly into all move causes high programming efforts and does not exploit structural dependen¬ cies that exist in most real- world transportation scenarios.

Therefore, the approach according to the invention is proposed based on preprocessing steps in the optimizer system, which reduce the model based on above four incompatibility classes to the core of the represented optimization problem. In this way, the problem stated as described above is transformed into an equivalent transportation optimization problem, which is modeled in a more compact fashion and in particular allows direct application of existing transportation optimizer software. This solution has the advantage that the internal search logic of the optimizer (i.e. the moves) is not affected at all and that the reduction already exploits all similarities and symmetries between order attributes and incompatibil¬ ity constraints. Therefore a significantly lower programming effort is needed, as well as a concentration of the search process on the core of the optimization problem. The latter would cause additional efforts in the straightforward approach discussed before.

The reduction step consists of several phases, which will be described in more detail. Fig. 4 gives an overview of the involved modeling layers and the phases that transform one layer into the subsequent one.

The phases have the following purpose:

Phase 1 shrinks the relational data. All attributes and values associated to each order are replaced by a single value for each order. This value is called group. As a result of the transformation, all orders with same group value have identical order characteristics, i.e. they have the same attributes and the same values of these attributes. The resulting model is more compact than the original input model for optimization.

Phase 2, which is optional, builds an object-oriented model, which is used for the internal optimizer search routines. It allows fast navigation between related objects (e.g. vehicles and orders) and quick access to the orders' properties (e.g. the group is directly accessible for each order). It contains some preprocessing steps that reduce the search efforts.

Phase 3 further reduces the incompatibility matrix, which stores for each pair of groups whether they are compatible or incompatible. Basically, this reduction replaces groups by equivalence classes of groups, where two groups are perceived as equivalent if they have the same incompatibility relations to all other groups. As a result, each order is associated to a reduced group. The number of reduced groups may be much smaller than the number of groups. This reduction reduces the whole incompatibility structure among groups to the core incompatibility structure, which cannot be further reduced by any technique.

Next, the four models and the three phases are described in more detail.

The first phase transforms the input model for optimization to the compact input model. Fig. 5 gives an overview. The input model for optimization is the output of the compatibil¬ ity engine, enhanced by a table which specifies the order characteristics. The transportation optimizer of course gets lots of additional tables, like distance matrices, vehicle costs, time windows, etc., but here we describe only those involved in the processing of incompatibili¬ ties. The model consists of tables, each of which can be processed by looping over all rows contained. The compact input model basically consists of tables, but all characteristics as¬ sociated to the orders are replaced by a single value associated to each order. This value is called group. All those orders that have exactly identical characteristics (i.e. the same set of attributes and the same values for each attribute) are assigned to the same group. Thus, a group represents a cluster of orders with identical properties.

The second phase transforms the compact input model to the object-oriented optimizer model, as illustrated by Fig. 6. The object-oriented optimizer model is used in the search process of the optimizer. It is designed to allow fast access to properties of optimization objects like e.g. orders and vehicles. Furthermore, it allows navigation between related objects, e.g. all feasible vehicles are directly reachable from a given order. The incompati¬ bility matrix allows determining whether two groups are compatible in a single step. For each order, the feasible legs are calculated. Among all possible legs those that violate in¬ compatibility constraints are already filtered out. For each leg, all feasible vehicles are de¬ termined, also based on the incompatibilities declared.

The third phase transforms the object-oriented optimizer model to the reduced object- oriented optimizer model, as illustrated by Fig. 7. The object-oriented optimizer model is already suitable for the optimizer, i.e. it could be used to search for good transportation plans. However, there is still the possibility to reduce the model considerably. Therefore, the third phase analyzes the incompatibility matrix and detects equivalence classes among the groups. Then, all original groups in the object-oriented model are replaced by reduced groups (abbreviated by RGroup). This may yield a substantial reduction of the incompati¬ bility matrix. This supports the optimizer's search process, since all orders, which are equivalent regarding their incompatibilities to other orders, now belong to the same (re¬ duced) group. Thus, the optimizer can detect compatibility for all equivalent orders by simply comparing their RGroup. Without this reduction, this check would always be done by looking up in the incompatibility matrix, which is fast, but not as fast as directly com¬ paring the RGroup attribute of the order objects.

In the following, the detailed design is described. Data types can be adjusted taking into account attribute identifiers, values, condition identifiers appropriate for a particular im- plementation. The only requirement is that the data type length should be wide enough to fit to maximum attribute value. The following notations are used:

it_<name> - input table; et_<name> - output table; lt_<name> - internal (local) table; ls_<name> - work area of table with name <name>, or structure Iv <name> - local variable Definition of compatibility types - The existing TPVS field catalogue functionality is the basis for compatibility type definitions. Field catalogue contains attributes, which are related to planning objects, such as order, vehicle and cross-docking location.

The concept of compatibility type was introduced to represent possible (incompatibility requirement for transportation scenarios. It's mainly consists of two attributes from field catalogue, which represent objects to be compatible or not.

Compatibility type database structure is as follows:

ΛTTRJD1 and ΛTTRJD2 fields are linked to TPVS field catalogue database table, field ATTR_ID through foreign key relationship. Therefore prerequisite of compatibility type definition is declaration of corresponding attributes in field catalogue.

Compatibility type's definition is part of customizing. During particular customer imple¬ mentation projects, all possible compatibility types should be identified and defined.

Definition of compatibility rules - Compatibility rules (below - compatibilities) represent actual attribute values which identify possible (incompatible object combinations. Compatibility definition is based on defined compatibility types.

Compatibility database table structure is as follows:

ATTR_VALUE1, 2 contain specific attribute values, which are (incompatible (please re¬ fer to Example). The data type of these fields should be defined in such a way, that possi¬ ble attribute values can be stored and unambiguously converted to/from them. In our de- sign character type of length 40 was used, as it is sufficient to store any of order attribute values, relevant for APO, vehicle ID and cross-docking location numbers.

Compatibility flag COMP_FLAG can have values 'C = 'Compatible', T - 'Incompati¬ ble', initial flag means that particular entry should be skipped. 'Compatible' flag means that all other possible values of second attribute in compatibility type, which were not de¬ clared as 'Compatible', will be incompatible (this is done to accelerate maintenance).

Conditions (existing functionality) can be used to restrict validity of corresponding com¬ patibilities only for specific orders. Condition ID is unique identifier of corresponding condition definition, i.e. it is linked to condition header table through foreign key relation¬ ship.

Compatibility definition is part of planning process. In particular it is possible to define them directly during transportation planning.

Compatibility Engine — determination of compatibilities Interface - input data: Input data contains planning objects which are relevant for determination of compatibilities. In particular these are orders, vehicle (IDs) and cross- docking location (numbers) (also often called hubs). Order table contains entries in form of order number, which uniquely identifies particular order in planning system, and all transportation-planning-relevant data, such as source and destination, delivery dates, quantities, shipping information, etc. Any order field is referred below as order attribute. Vehicle IDs and cross-docking location numbers are also referred below as attributes.

Input tables have following structure:

Orders it order.

Vehicle itjtype:

Cross-docking location itjiubjoc:

Interface - output data - Output contains incompatible attribute values, which are classified to 4 classes relevant for transportation, plus order attributes (in form of order number- attribute-ID-value combinations).

Output tables have following structure: Incompatible vehicle-order attributes etjncjtype ordattr:

Incompatible order attribute-order attributes et_inc_ordattr ordattr:

Incompatible cross-docking location order attributes et_inc_hub_ordattr.

Incompatible vehicle-cross-docking location etjnc_ttype_hub:

Order attributes et ord attr:

Field name Type(length) Comment

ORDNO CharlO Order number

ATTR ID CharlO Order Attribute ID

ATTR VALUE Char40 Order Attribute Value Important internal tables - Table of compatibility types lt_comp_types (the structure is identical to database table structure):

Table of compatibility entries lt_comp_all, for specific compatibility type lt comp (the structure is similar to database table structure, ID and planning profile fields are omitted for simplicity, as they are not used):

Attribute ID table It attribute ids:

Table of attribute ID-value combinations for input objects lt_attr_id_yalues:

Table of incompatible combinations of attribute ID- values lt_inc_attrjd_yalues:

In the following, an algorithm of determination of incompatibilities is described. Reference is made to Figs. 8-13. In these figures, compatibility entries are denoted as 'rules' for brevity.

1. Read compatibility types lt_comp_types, compatibility rules lt_comp_all from cor¬ responding database tables for current planning profile.

At this step relevant compatibility types and compatibility rules should be read and passed to internal tables with structure identical with customizing database table structures. Table of used attribute IDs lt_attribute_ids is filled by looping over compatibility types lt_comp_types and adding attribute Ids, ATTR IDl and ATTR ID2 to it.

2. Prepare attribute ID-value combinations for input objects lt_attr_id_yalues (see Fig.

9). At this step combinations of attribute IDs and values are prepared for input objects in form of attribute ID - value pairs.

2.1. Loop at table of attribute IDs lt_attribute_ids into lv_attribute_id; 2.1.1. determine, what object attribute ID refers to (order, vehicle, cross-docking loca¬ tion). In our design the reference is hard-coded, in particular, attribute with ID = "TTYPE' always refers to vehicle, 'HUB_LOC - to cross-docking location, any other attribute - to order attribute. 2.1.2. Case object type:

2.1.2.1. vehicle - loop at vehicle table itjtype into Isjtype and for each entry add entry with

ATTRJD = Ivjittributejd ( 'TTYPE') ATTR_VALUE = isjtype-ttype;

2.1.2.2. cross-docking location - loop at cross-docking location input table it_hub_loc into Isjiubjoc and for each entry add entry with ATTRJD = Ivjittributejd ('HUB_LOC'),

ATTR_VALUE = is Jiub Joe-hub Joe;

2.1.2.3. order - loop at order input table it orders, for each entry get value of attribute by:

2.1.1.3.1. calling user exit 2.1.1.3.2. accessing order structure using component name = Ivjittributejd as refer¬ ence, if user exit is not active

2.1.1.3.3. if failed, add error message to log and process next order

2.1.1.3.4. add entry with ATTRJD = Ivjittributejd, ATTR_VALUE = returned value of order attribute

3. Loop at compatibility types lt_compjypes into ls_compjype.

3.1. Prepare compatibility entries of current type lt_comp using table lt_comp_all.

3.2. Prepare incompatible combinations of attribute values (see Fig. 10). Loop at compatibility entries of current type lt_comp into ls_ comp.

Check if entry in table lt_attrjd_yalues with key ls_comp -ATTRJDl, ls_comp-ATTRJ^rALUEl exists; If FALSE, process next compatibility entry; 3.2.1. Case compatibility flag (ls_comp-COMP_FLAG): 3.2.1.1. 'C (Compatible).

In this case it is necessary enter all other attribute values as incompatible, which are not explicitly declared as compatible. 3.2.1.1.1. Loop at lt_attr_id_values into ls_attrjd_yalues where ATTRJD = Is_compjype-ATTRJD2.

3.2.1.1.1.1. Check if ls_attr_id_values-ATTR_VALUE = ls_comp-ATTR_VALUE2 (same rule). IfTRUE, process next compatibility entry.

3.2.1.1.1.2. It is necessary to check if particular attribute value is entered as compatible in another compatibility entry with the same condition or without condition. For entry with condition it is necessary to check another entries either with same condition ID, or with initial condition ID. For entry without condition, it is necessary to check another entries with any condition or without it.

3.2.1.1.1.3. Check if NOT ls_comp-CONDITION_ID is initial.

3.2.1.1.1.4. If TRUE

3.2.1.1.1.4.1. Check entry in table lt_comp with key

ATTR_VALUE1 = ls_comp-ATTR_VALUEl ATTR VALUE2 = Is_attrjd_values-ATTR_VALUE

COMP FLAG = 'C

CONDITIONJD = ls_comp-CONDITIONJD exists.

If TRUE, process next compatibility entry. 3.2.1.1.1.4.2. Check entry in table lt_comp withkey

ATTR VALUEl = ls_comp-ATTR_VALUEl

ATTRJAIΛJE2 = Is_attrjd_values-ATTR_VALUE

COMPJF⁷LAG = 'C

CONDITIONJD = 0 exists.

If TRUE, process next compatibility entry.

3.2.1.1.1.5. ELSE

3.2.1.1.1.5.1. check entry in table lt_comp with key

ATTR JALlJEl = ls_comp-ATTR_VALUEl ATTR_VALUE2 = Is_attrjd_values-ATTR_VALUE

COMPJFLAG = V exists. IfTRUE, process next compatibility entry.

3.2.1.1.1.6. Check entry in table lt_comp with key

ATTR JALUEl = ls_comp-ATTR_VALUEl ATTR_VALUE2 = Is_attrjd_values-ATTR_VALUE COMPJ¹LAG = 'C

CONDITIONJD = 0 exists.

IfTRUE, process next compatibility entry.

3.2.1.1.1.7. Add incompatible combination entry to table lt_inc_attr_id_yalues:

ATTRJDl = ls_comp-ATTRJDl ATTR JALUEl = Is _comp-ATTR JALUEl ATTR JD2 = h_comp-ATTRJD2 ATTRJALUE2 = Is_attrjd_values-ATTR JALUE CONDITIONJD = ls_comp-CONDITIONJD

3.2.1.1.2. End loop at lt_attrjd_values.

3.2.1.2. T (incompatible) Append incompatible combination entry to table ltjnc_attrjd_yalues:

ATTRJDl = ls_comp-ATTRJDl

ATTR JALUEl = Is _comp- ATTR JALUEl

ATTRJD2 = ls_ comp-ATTRJD2 ATTRJALUE2 = Is _comp- ATTR JALUE

CONDITIONJD = ls_comp-CONDITIONJD

3.2.1.3. Initial - process next compatibility entry.

3.3. End loop at lt_comp.

3.4. Process incompatibilities with conditions (see Fig. 11). 3.4.1. Loop at incompatible attributes of current type lt_inc_atfr_id_values into ls_inc_attrjd_values where CONDITIONJD is NOT initial

3.4.1.1. Append Isjnc_attrjd_values-CONDIΗON_ID to lt_condition_ids. 3.4.2. End loop.

3.4.3. Read assignments of orders to conditions lt_order_to_condition (external function of condition processing).

3.4.4. Loop at lt_condition_ids into ls_condition_id.

For each condition a unique virtual attribute ID and value for order is generated. If order satisfies condition, then it has corresponding virtual attribute ID-value combination. In¬ compatibilities with conditions are converted to incompatibilities between virtual attrib¬ utes.

3.4.4.1. Generate new attribute ID Ivjiew attributejdl for attribute 1 of current com¬ patibility type ls_comp_type-ATTR_IDl, if attribute 1 refers to order (see 2.1.1.).

3.4.4.2. Generate new attribute ED Iv_new_attribute_id2 for attribute 2 of current com¬ patibility type ls_comp_type-ATTR_ID2, if attribute 1 refers to order (see 2.1.1.).

In our design, new attribute ID is constructed by concatenation of attribute ID and condi¬ tion ID (see example).

3.4.4.3. Loop at incompatible attributes of current type lt_Jnc_attr_id_values into ls_inc_attr_id_values where CONDITIONJD = ls_condition_id. 3.4.4.3.1. Generate new attribute value lv_new_attribute_valuel for ls_inc_attr_id_values-ATTR_VALUEl, if attribute 1 of current compatibility type refers to order.

In our design, new attribute value is constructed by concatenation of attribute value and condition ID (see example). 3.4.4.3.2. Replace entry Isjncjzttr _id_values-ATTR_VALUEl in ltj,nc_attr_ idjyalues with new value lv_new_attribute_yaluel 3.4.4.3.3. Append generated attribute for orders which satisfy current condition: 3.4.4.3.3.1. Loop at lt_orderjo_condition into ls_orderjo_condition where condi¬ tion Jd = ls_condition_id.

3.4.4.3.3.1.1. Append order attribute to table et_ord_attr:

ORDNO = ls_order to_condition-ORDNO ATTRJD = ls_new_attributejdl

ATTR_VALUE = lv_new_attribute_valuel if attribute 2 of current compatibility type refers to order.

3.4.4.3.3.2. End loop at lt_orderjo_condition.

3.4.4.3.4. Generate new attribute value Iv_new_attribute_yalue2 for Isjnc_attrjd_values-ATTR_VALUE2, if attribute 2 of current compatibility type refers to order.

In our design, new attribute value is constructed by concatenation of attribute value and condition ID (see example). 3.4.4.3.5. Replace entry Isjnc_attrjd_values-ATTR_VALUE2 in ltjnc_attrjd_yalues with new value Iv_new_attήbute_yalue2 3.4.4.3.6. Add generated attribute for orders which satisfy current condition. 3.4.4.3.6.1. Loop at lt_orderjo_condition into ls_orderjo_condition where condi¬ tion Jd = ls_conditionjd. 3.4.4.3.6.1.1. Append order attribute to table et_ord_attr:

ORDNO = Is_orderjo_condition-ORDNO, ATTRJD = Iv_new_attributejd2, ATTRJfALUE = Iv_new_attribute_value2, if attribute 2 of current compatibility type refers to order. 3.4.4.3.6.2. End loop at lt_orderjo_condition.

3.4.4.4. End loop at ltjnc_attrjd_values.

3.4.5. End loop at lt_conditionjds.

3.5. Classify incompatibilities into 4 classes (see Fig. 12). 3.5.1. Loop at ltjnc_attr_id_yalues into ls_inc_attr_id_values.

3.5.1.1. Determine, what object attributes 1 and 2 refer to (order, vehicle, cross- docking location) (see 2.1.1.).

3.5.1.2. Case ATTRJDl -ATTRJD2 (object combination): 3.5.1.2.1. Vehicle — Order attribute, or Order Attribute - Vehicle. Append entry to etjnc_ttype_prdattr:

TTYPE = ls_inc_attrJd_yalues-ATTR_yALUEl ATTR JD = Isjnc_attrjd_values-ATTRJD2 ATTR_VALUE = Isjnc_attrjd_values-ATTR_VALUE2 3.5.1.2.2. Order attribute - Order attribute. Append entry to etjnc_ordattr_ordattr. ATTR JDl = Isjnc_attrjd_values-ATTRJD1 ATTR_VALUE1 = Isjnc_attrjd_values-ATTR_VALUE1 ATTR JD2 = Isjnc_attr id_values-ATTRJD2 ATTR VALUE2 = Isjnc_attrjd_values-ATTR_VALUE2 3.5.1.2.3. cross-docking location - Order attribute, or Order Attribute - cross-docking location. Append entry to etjncjιub_ordattr. HUBJOC = Isjnc_attrjd_values-ATTR_VALUE1 ATTR ID = Isjnc_attrjd_values-ATTRJD2 ATTR_VALUE = Isjnc_attrjd_values-ATTR_VALUE2 3.5.1.2.4. Vehicle - cross-docking location, or cross-docking location — vehicle. Ap¬ pend entry to etjncjtypejtub: TTYPE = Lsjnc_attrjd_values-ATTR_VALUE1 HUBJOC = Isjnc_attrjd_values-ATTR_VALUE2

3.5.1.2.5. Other combination - error message 'Wrong compatibility type'.

3.5.2. End loop at ltjnc_attrjdjyalues.

3.5.3. Filter out not-used order attributes.

3.5.3.1. Loop at et_ord_attr into ls_ord_attr.

3.5.3.1.1. Read table ltjnc_attrjd_yalues with key ATTRJDl = Is ord_attr-ATTR ID, ATTR J ^r AUJEl = ls_prd_attr-VALUE. If entry found, process next ls_ord_attr.

3.5.3.1.2. Read table lt_inc_attr_id_yalues with key

ATTRJD2 = ls_ord_attr-ATTRJD, ATTR_VALUE2 = ls_ ord_attr-VALUE. If entry found, process next ls_prd_attr.

3.5.3.1.3. Delete current order attribute from et ord attr.

3.5.3.2. End loop at et_ord_attr.

4. End loop at lt_comp_types.

Optimizer Preprocessing of Incompatibilities - In the following, the terms order characteristic and order attribute will be used as synonyms.

Characteristics-based clustering of orders into groups - The main task is to determine orders which have identical characteristics, and put them together into a group. Basically, the key input for this phase is the table that declares the orders' characteristics. Due to a mapping of strings to consecutive integers, the following (numeric) structure is assumed:

Orders, characteristics and values are numbered consecutively, starting by 1. A row of this table has the semantic that the order with ID ORDER has the value VAL regarding its at¬ tribute CHAR.

The following algorithm determines clusters of orders. A cluster is a set of orders, which have (i) the same set of attributes, and (ii) the same value for each declared attribute, re- spectively. We use the following notation: M : A -> B is a mapping from A to B. |M| is the number of elements from A that are mapped to B.

Algorithm Clustering:

Let dust : (char -> val) -> orders be a mapping that maps mappings (char -> val) to a list of orders. Initially, dust is empty, i.e. \dust\ = 0.

For each order do begin

Let m: char -> val be a mapping that is empty, i.e. \m\ = 0. For each (char,val) declared for the order do Set m(char) = val.

If dust already contains m then

Append order to clust(m). else

Set clust(m) = (order). end

Set counter— 1.

For all mappings (char -> val) in dust do

Assign all orders in the list clust(char->vaϊ) to the group counter.

Assign the group 0 to all orders not assigned to any group in previous loop.

End Algorithm Clustering: It should be noted that this is high-level pseudo-code. Its purpose is to clarify the basic idea. Of course, several additional maps may be required to implement this algorithm effi¬ ciently. In particular, processing of orders requires several inverse maps that determine for a given characteristic and value all corresponding groups. All the intermediate results of this algorithm are stored in maps, which can be efficiently accessed by the next transfor¬ mation steps. The two loops listed in the algorithm can be unified into a single loop, which is even more efficient.

In the exemplary implementation of the algorithm, several checks of the input data are in- eluded, which immediately report errors or inconsistencies in the input data, like e.g. the inconsistent declaration of several values for one characteristics of one order.

Build Object-Oriented Optimizer Model — The core optimizer engine of our TPO uses an object-oriented model, where the different objects are linked and the information needed is better accessible than in a relational-table based model.

Four objects are relevant in connection with (incompatibilities:

- The OrderObject stores the OrderGroup computed in the previous step when clustering the order characteristics with their values into compact groups. Later in the process this object also stores the reduced group determined during building equivalence classes. Additionally, the OrderObject of course stores all other relevant information of the orders contained in other input tables, which aren't relevant for (incompatibilities. - The second object is the incompatibility matrix which is another representation of the table of incompatibilities between OrderGroups.

- The other two objects, the FeasibleLegsOfOrder and Feasible VehiclesOfLeg combine information of (in)compatibilities between vehicle and orders, orders and hubs and hubs and vehicle, using the hub structure and available vehicle of the underlying opti- mization problem (given in some of the other input tables). The (feasible) legs of an order depend on the hub-network of the problem. We can have a 'direct leg' connecting the pickup location (depot, plant, etc) directly with the delivery location (customer, etc) of the order. If hubs are available, there is the possibility to reload the order at these cross-docking locations. There exist legs from the pickup location to the reachable hubs, legs from some of the hubs to the delivery location and of course there may exist some legs between some of the hubs.

Entries in the incompatibilities table between orders and hubs lead directly to the elimina¬ tion of some of the legs, namely of those where one of the two locations is an incompatible hubs for the order.

Each of the leg can be served by some vehicles; some vehicles may not reach some hubs because of the corresponding distance matrix (e.g. a ship can only serve some harbors but no airports, the distance from the harbors to the airport were set to infinite). Additionally some incompatibilities between hubs and vehicle may be maintained in the corresponding table, which lead to a further elimination of vehicle at the legs where the specified hub is involved.

If some vehicles are not compatible to a specific order, these vehicles are not usable at any of the legs of the order, which means that they must be deleted from all FeasibleVehicle- sOfLeg-Objects of the feasible legs of the order. As a result, for some legs no feasible ve¬ hicle may be left, which causes the infeasibility of the leg. In such case, the leg can be re¬ moved from the set of FeasibleLegsOfOrder. Moreover, all legs which can't be extended with any other feasible legs to an edge path from the pickup to the delivery location of the order can also be removed, since they cannot be used to perform the order.

An example of the transformation to the object-oriented optimization model is given in the example section for optimizer preprocessing. Build Reduced Groups - It may happen that some of the OrderGroups are in the same equivalence class regarding incompatibilities with other OrderGroups, e.g. some groups are compatible with all other groups, or one OrderGroup is incompatible to several other groups which have no fiirther incompatibilities. Some other OrderGroups may not be in use since e.g. the orders using the OrderGroups cannot be performed at all or in the case of decomposing the problem into smaller subproblems the corresponding orders are not selected to be optimized in the current decomposed subproblem. Building these equivalence classes helps to reduce the size of the incompatibility matrix, which leads to a better performance when checking for incompatibility between orders and other orders or hubs or vehicle.

We assume an incompatibility matrix with entries 'compatible' and 'incompatible' for each of the order groups. The incompatibility matrix has two important characteristics:

- for all OrderGroups g: g compatible g symmetry (gl compatible g2 if and only if g2 compatible gl)

In the further search only those OrderGroups must be considered, which are present in the current subproblem in at least one performable order. The incompatibility matrix can there- fore be reduced by eliminating the unnecessary OrderGroups, which can be done by elimi¬ nating the corresponding rows and columns from the incompatibility matrix.

The next step is to detect OrderGroups with identical incompatibilities, which belong to the same equivalence class and can be treated as one single virtual incompatibility group (fur- ther called 'RGroup', abbreviated from ReducedGroup). To detect the equivalence classes one has to compare the reduced rows (or columns because of the symmetry) of the incom¬ patibility matrix. Each two identical rows belong to the same equivalence class, because they have the same incompatible OrderGroups. Each equivalence class defines an RGroup. The incompatibility matrix can now be reduced furthermore. For each of the equivalence classes, one reference member of the group with its corresponding row and column will remain in the incompatibility matrix, while the rows and columns for the other members of the RGroup will be deleted. The remaining incompatibility matrix (with at least one row and column) is the core in¬ compatibility matrix, which can't be reduced any more. Every two rows (and columns) are pair wise different.

Finally the OrderGroups in the orders have to be updated / replaced by the mapped RGroups.

Algorithm for the reduction (not including updates of OrderGroups):

input:

- incompatibility matrix of the OrderGroups

- selected performable orders output: - mapping of OrderGroups to RGroups

- incompatibility matrix of the RGroups

begin

Eliminate not used OrderGroups in the matrix by deleting corresponding rows and col- umns

Build equivalence classes (rows (and columns) which are identical belong to the same equivalence class)

Assign new RGroup to each equivalence class Shrink incompatibility matrix according to equivalence classes end In the following, an example of compatibility definitions and processing is given. A typical transportation scenario includes following requirements:

1. Orders of certain transport groups should not be combined onto the same shipment. In particular, transport group FOOD is not compatible with ELECTRONICS.

2. Certain vehicle should be used for orders of specific shipping conditions. I.e. vehicle FEDEX should be used for orders of shipping conditions 01, for orders of shipping conditions 02 vehicle DHL may be used, if order destination is USΛ, TNT - if order destination is EUROPE, for other destinations there are no specific requirements.

Customizing - Field catalogue contains following fields:

The following compatibility types are created: - 1 transport group - transport group

2 shipping conditions - vehicle.

In this case Compatibility Types database contains following entries:

Setup of compatibility rules Following compatibility rules are created (I = incompatible, C = compatible):

Here condition USA corresponds to orders with destination USA, condition EUR corre¬ sponds to orders with destination EUROPE.

Determination of incompatibilities

Input data - Suppose that there are following orders and vehicle during transportation planning:

Order data:

Vehicle data:

Contents of important internal tables at processing steps described in process diagram.

Attribute ID table It attribute ids:

ATTR ID

TRAGR VSBED TTYPE

Table of attribute ID-value combinations for input objects lt_attr_id_yalues:

Step 1 of loop at compatibility types, compatibility type = 1.

Table lt_inc__attr_id_values of incompatible combinations of attribute ID -values.

Condition processing is not relevant, as they are not defined for compatibility rules of compatibility type 1.

Tables of incompatibilities after classification of incompatible entries et inc ordattr ordattr:

Order attributes to be appended to et_ord_attr.

Step 2 of loop at compatibility types, compatibility type = 2.

VSBED 02 TTYPE FEDEX EUR

VSBED 02 TTYPE DHL EUR

VSBED 02 TTYPE UPS EUR

It condition ids:

Condition ID

USA EUR

Table It order to condition:

Table lt_inc_attr_id_yalues — after condition processing (here the unique identifiers for attributes with conditions are constructed by concatenation of attribute ID/value with con¬ dition ID):

VSBED_EUR 02_EUR TTYPE DHL EUR VSBED EUR 02 EUR TTYPE UPS EUR

Tables after classification of incompatible entries:

Incompatible vehicle-order attribute:

Order attributes (before filtering of unused attributes):

Order attributes to be appended to et_ord_attr (after filtering of not-used attributes):

Output tables

Incompatible vehicle-order attributes:

UPS VSBED EUR 02 EUR

Order attributes:

Other tables for incompatibilities are empty. Example for optimizer processing of incompatibilities

Characteristics-based clustering of orders into groups - Here, we focus on the table that defines the characteristics of the different orders. An example table with the three columns ORDER, CHAR, and VAL is given below:

The clustering algorithm produces a result, which can is shown from two different perspec¬ tives:

The first perspective is the list of all clusters. For each group, the characteristics and values are given, as well as the orders that have exactly these characteristics and values.

The second perspective lists for each combination of characteristic and value, for which at least one order exists, the clusters that contain the orders having this com¬ bination.

Now, here is the outcome of the algorithm, viewed from both perspectives.

There are ten groups. The first nine were explicitly generated by the first loops in the clus¬ tering algorithm, whereas the tenth contains those orders which are not contained^' in the first nine groups. This last group contains orders (in our example 5 and 15), for which no characteristics were declared in the input table.

Build Object-Oriented Optimizer Model - Assume the hub network ('Hubl' - 'Hub3') with the possible legs ('Ll' - 'L9') for one selected order (has OrderGroup 'GroupF) illustrated in Fig. 13.

Now assume table IncHubGroup with the entries

Hub3 must be removed including all legs where Hub3 is involved (these are the legs 'L7' - 'L9'). The remaining legs are feasible legs of the order, see Fig. 14.

The general assignment (without considering further incompatibilities) of vehicle to the remaining feasible legs is shown in the following table:

The next step is to process table IncVehHub, which in this example looks like

As a result, the legs L2, L5 and L6 (Hub2 is one location of the leg) can't use V2, this ve¬ hicle has to be removed from the set of feasible vehicles.

If we additionally assume incompatibilities between vehicle and orders (OrderGroups, ta¬ ble IncVehGroup), we must reduce the feasible vehicles of a leg.

Vl has to be removed from all legs. The final Feasible VehiclesOfLeg-objects were filled as written below:

Leg Vehicle

Ll V2, V3

L2 V3, V4

L3 V2, V3, V4

L4 V2, V3, V4

L5 V3, V4, V5

L6

Leg L6 has no feasible vehicle anymore, so it cannot be used, see Fig. 15.

As no path from the pickup location to delivery location including legs L2 and L5 can be built, these two legs are not usable and they are removed from the model. The remaining FeasibleLegsOfOrder objects are the objects for Ll, L3 and L4 with its Feasible Vehicle- sOfLeg objects, see Fig. 16.

Build Reduced Groups - The following given orders with their OrderGroups are assumed:

The incompatibility matrix resulting from clustering incompatibilities into OrderGxoups looks as follows:

The OrderGroups 9 and 13 can be eliminated since the orders using those OrderGroups (orders O9 and 014) are not selected or not performable, but OrderGroup 6 must remain, since only one of the two orders using this OrderGroup is not selected (order O6), while the other order O7 still can be performed in the current subproblem.

After removing the corresponding rows and columns from the incompatibility matrix, it looks as follows:

Now one can find the equivalence classes, which are:

Shrinking the incompatibility matrix to one reference member per equivalence class (delete the rows and columns of all but one member of the equivalence classes) yields the incom¬ patibility matrix

This matrix is (with replacing the reference OrderGroup per equivalence class with the RGroup assigned to the equivalence class) the final reduced incompatibility matrix:

To complete the output, this is the mapping of OrderGroups to RGroups:

It is assumed that transportation optimizer is capable to respect four "atomic" incompatibil¬ ity constraints, which are retrieved by compatibility engine:

Group-Group: Order group Λ is incompatible with order group B, i.e. it is not al¬ lowed to transport orders from group Λ on the same vehicle as orders from group B.

- Group-Hub: Order group A must not be shipped indirectly through hub B . - Group- Vehicle: Order group A must not be shipped by vehicle B.

Vehicle-Hub: Vehicle A must not load/unload orders at hub B.

(Cross-docking loc ation corresponds to hub, group is order characteristic in optimizer model).

Therefore the incompatibilities, which were determined by compatibility engine, can be directly mapped to corresponding optimizer tables.

Manual planning is direct assignment of specific freight units to vehicle of particular vehi- cle.

To detect and inform planner about possible incompatibilities the relevant objects can be collected and passed to compatibility engine. The engine provides as output incompatibili¬ ties which were encountered between particular objects. Therefore they can be displayed to alert user.

Fig. 17 is a screen shot of a Compatibility Type definition transaction screen.

Figs. 18 and 19 are screen shots of Compatibility definition transaction screens. The present techniques can be implemented in digital electronic circuitry, or in computer hardware, firmware, software, or in combinations of them. Apparatus of the invention can be implemented in a computer program product tangibly embodied in a machine-readable storage device for execution by a programmable processor. Method steps according to the invention can be performed by a programmable processor executing a program of instruc¬ tions to perform functions of the invention by operating on the basis of input data, and by generating output data. The invention may be implemented in one or several computer programs that are executable in a programmable system, which includes at least one pro- grammable processor coupled to receive data from, and transmit data to, a storage system, at least one input device, and at least one output device, respectively. Computer programs may be implemented in a high-level or object-oriented programming language, and/or in assembly or machine code. The language or code can be a compiled or interpreted lan¬ guage or code. Processors may include general and special purpose microprocessors. A processor receives instructions and data from memories, in particular from read-only memories and/ or random access memories. A computer may include one or more mass storage devices for storing data; such devices may include magnetic disks, such as internal hard disks and removable disks; magneto-optical disks; and optical disks. Storage devices suitable for tangibly embodying computer program instructions and data include all forms of non-volatile memory, including by way of example semiconductor memory devices, such as EPROM, EEPROM, and flash memory devices; magnetic disks such as internal hard disks and removable disks; magneto-optical disks; and CD-ROM disks. Any of the foregoing can be supplemented by or incorporated in ASICs (application-specific inte¬ grated circuits).

The computer systems or distributed computer networks as mentioned above may be used, for example, for producing goods, delivering parts for assembling products, controlling technical or economical processes, or implementing telecommunication activities.

To provide for interaction with a user, the invention can be implemented on a computer system having a display device such as a monitor or LCD screen for displaying informa¬ tion to the user and a keyboard and a pointing device such as a mouse or a trackball by which the user can provide input to the computer system. The computer system can be programmed to provide a graphical or text user interface through which computer pro¬ grams interact with users.

A computer may include a processor, memory coupled to the processor, a hard drive con¬ troller, a video controller and an input/output controller coupled to the processor by a proc¬ essor bus. The hard drive controller is coupled to a hard disk drive suitable for storing ex¬ ecutable computer programs, including programs embodying the present technique. The I/O controller is coupled by means of an LO bus to an VO interface. The I/O interface re- ceives and transmits in analogue or digital form over at least one communication link. Such a communication link may be a serial link, a parallel link, local area network, or wireless link (e.g. an RP communication link). A display is coupled to an interface, which is coupled to an I/O bus. A keyboard and pointing device are also coupled to the I/O bus. Alternatively, separate buses may be used for the keyboard pointing device and I/O inter- face.

Other embodiments are in the scope of the following claims.

Claims

Claims

1. A method of modeling incompatibilities in a transportation planning and optimiza- tion system, wherein a transportation plan is determined for

a plurality of orders (1), each order (1) being defined by a predefined num¬ ber of characteristics, one of the characteristics specifying freight units (10) to be transported from a predetermined place of departure to a predeter- mined place of destination, each freight unit representing a predefined good, a plurality of vehicles (20) for transporting freight units (10), and a plurality of cross-docking locations (30) for loading/unloading freight units (10) onto/from vehicles (20),

the transportation plan defining shipments, each shipment specifying the vehicles

(20), and cross-docking locations (30) needed for fulfilling the plurality of orders (1), whereby a number of freight units (10), vehicles (20), and cross-docking loca¬ tions (30) are incompatible to each other such that they are not allowed within the same shipment;

each of the freight units (10), vehicles (20), and cross-docking locations (30) being specified by attributes which are comprised in a field catalogue (100);

the modeling method comprising the following steps:

providing a data structure for the field catalogue (100), the field catalogue (100) providing data field attributes, each of the attributes representing one of

- freight units (10),

- vehicles (20), and - cross-docking locations (30), respectively; providing a data structure for a plurality of compatibility types (200), the data struc¬ ture providing for each compatibility type data fields for:

- a first entry (210) for linking the compatibility type structure with a first attribute of the field catalogue (100);

- a second entry (220) for linking the compatibility type structure with a second attribute of the field catalogue (100);

providing a number of different compatibility types (200), each compatibility type (200) being described by values of the data fields as defined in the provided com¬ patibility type data structure;

whereby a number of compatibility rules (300) is provided, each compatibility rule (300) specifying a combination of two values which refer to first and second attrib- utes of a defined compatibility type (200), and being indicative whether or not they are compatible,

the compatibility rules (300) being applied upon determining compatible combina¬ tions of freight units (10), vehicles (20), and cross-docking locations (30) for the shipments of the transportation plan.

2. The method of claim 1, wherein the compatibility type data structure further pro¬ vides a data field for an ID (230) for identifying the respective compatibility type (200).

3. The method of claim 1 or 2, wherein the compatibility type ID (230) is a numeric value.

4. The method of one of the preceding claims, wherein the compatibility types (200) comprise at least one of - transportation group, and transportation group;

- vehicle, and transportation group;

- incoterm, and incoterm;

- cross-docking location, and transportation group; - delivery priority, and cross-docking location;

- delivery priority, and vehicle;

- shipping conditions, and cross-docking location;

- shipping conditions, and vehicle;

- weight/volume group, and vehicle; - weight/volume group, and cross-docking location;

- transportation zone, and vehicle; and cross-docking location, and vehicle.

5. The method of one of the preceding claims, wherein a compatibility rule (300) fur- ther comprises a condition ID for restricting the applicability of the compatibility rule (300) onto predefined orders.

6. A method of processing incompatibilities for a transportation plan for

a plurality of orders (1), each order (1) being defined by a predefined num¬ ber of characteristics, one of the characteristics specifying freight units (10) to be transported from a predetermined place of departure to a predeter¬ mined place of destination, each freight unit representing a predefined good, a plurality of vehicles (20) for transporting freight units (10), and a plurality of cross-docking locations (30) for loading/unloading freight units (10) onto/from vehicles (20),

the transportation plan defining shipments, each shipment specifying the vehicles (20), and cross-docking locations (30) needed for fulfilling the plurality of orders, whereby a number of freight units (10), vehicles (20), and cross-docking locations

(30) are incompatible to each other such that they may not be allowed within the same shipment; each of the freight units (10), vehicles (20), and cross-docking locations (30) being specified by attributes which are comprised in a field catalogue (100);

the method comprising the following steps:

Receiving data which is descriptive of the orders (1), data which is descriptive of the available vehicles (20), and data which is descriptive of the available cross-docking locations (30);

reading a set of compatibility types (200) , and a set of compatibility rules (300) with respect to the plurality of orders (1), vehicles (20), and cross-docking locations (30), each compatibility rule (300) specifying a combination of two values which refer to first and second attributes of a defined compatibility type (200), and being indicative whether or not they are compatible;

associating a value with each order (1), the value being descriptive of the character¬ istics of the order (1), such that orders (1) having the same characteristics have the same value associated therewith;

replacing the data descriptive of the characteristics by the value associated therewith; and

outputting the order data resulting from the preceding step.

7. The method of claim 6, wherein prior to the outputting step, the following steps are performed:

classifying orders (1) having the same value associated therewith into a group; replacing the groups which have identical incompatibility relations to all other groups by a class; and

creating a data representation of the orders (1), whereby the orders (1) are grouped into classes of orders (1) having the same incompatibilities to the other classes.

8. The method of claim 6 or 7, further comprising:

Representing the groups of orders (1) by an object-oriented model, the model com- prising as one object the group of orders (1), and as second object an incompatibil¬ ity matrix, the matrix representing the incompatibilities between pairs of groups; as third object a data structure representing incompatibilities between vehicle (20) and orders (1), and between orders (1) and cross-docking locations (30); and as fourth object incompatibilities between cross-docking locations (30) and vehicles (20).

9. The method of claim 8, wherein

each order object comprises a data field which specifies the group; the incompatibilities matrix object comprises for each pair of order groups an information as to whether they are compatible or not; each order object comprises a set of legs referring to relevant compatible cross-docking locations (30); and leg objects which comprise a set of compatible vehicles (20).

10. A computer system for performing the method according to one of the preceding claims.

11. A computer-readable storage medium comprising program code for performing the method according to one of claims 1 to 9, when loaded into a computer system.