EP1500567B1 - Method for resolving conflicts in a trackbound transportation system - Google Patents

Abstract

A process resolves the problem of conflicting rights of way (S0, S1, S2) for tracked vehicles e.g. railway system. The system has a series of resolution strategies (M1, M2, M11, M12, M22) that are based on past experience and which are known to result in given outcomes, while taking into account a group of desirable factors (11). A quality factor is determined for each of the possible solutions before selecting the best choice (11).

Description

State of the art

The invention relates to a method for controlling the occupancy in a track-bound traffic system in case of occupancy conflict case or one technical conflict case, with possible alternative measures for resolution the conflict case and taken into account when determining a solution For each solution, a quality factor is determined, and the solution with the best quality factor is applied.

For track-bound traffic systems, in particular at Railway systems, conflicts can occur, for example when two trains enter the same track at the same time or the same Want to prove track vacancy section (occupancy conflict case) or because technical incompatibility between a train and the infrastructure (technical conflict). A technical conflict case, i. an incompatibility between train and track, for example, exists between an electric powered train and a non-electrified track. Such conflicts must be resolved before they become a dangerous situation, one Malfunction or lead to an accident. To resolve a conflict means measures, such as the redirection of a case of conflict train to apply before this conflict happens. Which Measures for which conflicts are applicable, for example, by Operating concept of the railway administration determined. A measure on Applying a train means the train run (schedule) of a conflict change the train involved so that the conflict is resolved. When a such action, for example, redirecting a train, is applied, a new situation arises, which typically generates consequential conflicts. The Resolving a conflict therefore means taking action in an orderly manner Order to resolve the original conflict and any subsequent conflicts that caused by the application of measures to dissolve.

There is often more than one for an original conflict and its consequent conflicts Solution. This stems from the fact that the solution to an original conflict applicable measures and the measures to resolve a Consequential conflict case can be combined. A solution for the resolution of a Conflict case is therefore defined as an ordered n-tuple (record) of Measures that together solve the original conflict and all subsequent conflicts. To determine the best solution, it is known to have a figure of merit for each To determine solution. This happens, for example, by calculating a Cost function that represents the quality of the solution. The solution with the lowest cost is the best solution. The cost of a solution is only determinable when the whole solution has been determined. To the best solution for any given conflict, all must be possible at all Solutions to this conflict (= solution space) must first have been calculated in order to To be able to compare costs. This procedure is useful and usable, when there are few solutions to a specific case of conflict. Exist however Many solutions to a specific conflict, then can be computing times arise that are so long that the conflict can not be resolved in time.

From US 5,794,172 is a method for moving objects through a Multi-way system, in particular a railway system, known. It is intended that orders can be placed. To have a schedule for such To create an order, marginal and secondary conditions must be observed. Taking these conditions into consideration will provide a solution with the best Cost factor sought. The solution is determined in a hierarchical system, wherein for problem solving an expert system is provided, the rules in the Solution noticed.

Object of the invention

The object of the invention is to provide a method with which the occupancy in a track bound traffic system in a occupancy conflict case quickly and can be controlled without much computational effort.

Subject of the invention

This object is achieved according to the invention by a method of the beginning in which a selection of the resolution of the conflict in question taken from an experience base and only these Measure (s) to be taken into account when identifying solutions. The selection of measures that should be applied to a conflict can be one or include several measures. The conflict can be a Original conflict or a consequential conflict. As only a selection of Action is used, the solution space is limited because less Solutions must be determined. In the simplest case be on one Conflict case applied the measures taken from the experience base, All measures lead to a conflict-free situation. For every measure, which leads to a subsequent situation, the quality of this solution is determined. The Measure leading to the best solution is applied.

The success of an applicable measure will be assessed and only those as promising measures of the experience basis for determining taken from the solution of a conflict. The success of a measure is the Frequency with which this measure is the best solution to conflicts under certain conflict circumstances (= conflict class)., This frequency can be combined with a weighting factor expressing how much the best solution is better than the average solution in solution space. Especially the quality of a solution can be determined by a cost function. Consequently The weighting factor can express how much the best solution is cheaper than the cost average in the solution room. However, there are other measures conceivable, with which the success of a measure can be evaluated. It is too note that the same action is taken in different conflict classes, i. at different conflict circumstances, can be different success. to Evaluation of the success of action usually needs the entire Solution space of a conflict, because only in this way can the best solution be determined can be. If only the measures classified as promising the Experience basis for determining the solution of a conflict case and its possibly. If consequential conflicts are removed, the solution space is reduced. This means that fewer solutions need to be calculated because the Calculation is based only on a selection of measures.

Leads a measure taken from the experience base to one Consequential conflict, is advantageously the resolution of the consequence conflict case likewise only a limited number of measures of the empirical basis taken. This method has the advantage that the number of too determining solutions is reduced by an original and / or Consequence conflict not all, but only a selection of measures applied or taken into account. When a conflict is resolved must be the appropriate one for the original or consequential conflict simply looked up in an experience base. By the procedure becomes Computing time saved, because the solution space is limited. At each Original and Consequential Conflicts are no longer all applicable measures considered. Good solutions are obtained if the experience base is statistically significant. The maximum required computing power of a Conflict resolution system can be reduced. Besides, a good one Solution can be found faster.

In a preferred variant of the method, the basis of experience in low-performing times of a conflict resolution system and / or after a Erstimplementierung of the process, in particular the first installation in the Conflict resolution system, built. The experience base is built up by complete solution spaces are determined and evaluated for a conflict become. Because of the experience base even in times of slowdown is built, the computer power in the conflict resolution system more evenly exploited or used. The experience base can thereby in low-power times, i. if in a track-bound Traffic system little real conflict cases occur, be extended. The bigger the experience base is, the faster a good resolution solution can be of a conflict. So it is almost a temporal Decoupling of learning and the application of the learned.

It is particularly preferred if conflict types and conflict circumstance classes be defined, which are assigned to a conflict class and the Conflict classes are stored in the experience base. The experience in the experience base is stored, is the connection between the Circumstances of a conflict (conflict situation) and the success of measures, used to resolve conflicts in this conflict circumstance can. The conflict type and a conflict circumstance class form together a conflict class. This combination of the two components comes in a railway system only once before. Every conflict can be unique be assigned to exactly one conflict class, since each conflict is also accurate associated with one of its components (conflict resource class, conflict type) can be. A type of conflict, for example, by the Railway administration. She does that from her operational point of view. An example of a conflict type is "train following conflict on the open road". One Conflict type is therefore the definition of the circumstances of a conflict from view the railway administration according to the operational concept. Every conflict has to can clearly be assigned to exactly one conflict type.

In a preferred embodiment of the method, everyone Conflict class the applicable, i. the measures eligible for Resolution of a conflict associated. In particular, it may be provided that a railway administration is determined by its operational concept, which conflict types exist. For each type of conflict is still set, like train 1 and train 2 are determined. In addition, the determined Railway administration for each type of conflict an unordered list of Measures on train 1 and a disorderly list of measures that are applicable to move 2 of a conflict of a conflict type. An applicable Action is a measure that, for a given type of conflict, is at one certain of the two moves involved in the conflict. For example, the measures "extend hold at train_1" and "hold extend at Zug_2 "two different measures.

In a particularly preferred process variant, the applicable Measures sorted within a conflict class. This allows one fast and good access to appropriate measures to solve a Conflict. In particular, it may be determined that only the first or first Measures within a conflict class to determine a solution to Resolution of a conflict.

Advantageously, conflicts are determined by the conflict resource class classified according to their formalized circumstances. This classification happens unaffected by the view of the railway administration.

Preferably, the classification is based on abstract place and Vehicle parameters. Classification becomes abstract information on the characterization of the trains involved in the conflict and the Infrastructure (conflict location) used. Become the abstraction of this information preferably abstract vehicle parameters and abstract location parameters used. For example, abstract vehicle parameters are the train priority Train Performance, Train Run Priority, Line Priority Priority (Line Recoverability Priority). Main criteria for line recovery are Reserve times in a train run. If a train has very little or no Having spare time in his schedule, he is prone to unrecoverable Delays. Because he can not spend any spare time to To catch up on delays. These vehicle parameters have sorting values how high, medium, low. For example, get train runs with little Reserve time a high value for the parameter line restore. For example, abstract location parameters are a detour potential sorting values such as high, medium, low.

It can be provided that for each abstract vehicle parameters and for each abstract location parameter, a conflict formalization parameter is used. Conflicting formalization parameters have comparative values such as "train 1 greater than train 2". Here is one Conflict circumstance class an ordered n-tuple (record) of all Konfliktformaliserungsparameter. The values of Conflict formalization parameters for a given conflict arise from the logical link of the abstract vehicle parameters involved in this conflict and the abstract location parameters of the Conflict place. Each conflict can clearly be exactly one confluence class be assigned. This means that each conflict is exactly one type of conflict and is assigned to a conflict circumstance class. Every conflict class consists of a conflict type and a conflict circumstance class. this means also that each conflict is associated with exactly one conflict class. there Several different conflicts can be assigned to the same conflict type and several (possibly other) conflicts with the same conflict circumstances assigned to the same conflict circumstance class.

Advantageously, the classification system is adaptive and adapts to Complexity of the transport system to which it is applied. The number The conflict classes for a railway system depends on the number of Combinations of abstract parameters used in this railroad system can happen. This number results from the heterogeneity of the Vehicle fleets (train service), timetables (train run priority, priority of Line restoration), the infrastructure (diversion potential, partly Train run priority) and the operating concept (train priority) exactly this Rail system.

It may be provided that any applicable measure of a Conflict class is assigned a success factor. The success factor can be Be a success counter. In this case, it is recorded, for example, which measure a Conflict class that was applied to a conflict of this conflict class best solution. The success counter associated with this action then becomes incremented. On the basis of the success counter, the measures of a conflict class be sorted.

It is particularly advantageous if a sorting key is determined via which the appropriate measures for a conflict to be resolved in the Experience base can be found. The sort key sorts the Measures of a conflict class ascending after their success among the Conflicting circumstances of this conflict class. This results in the best, the second best, the third best etc. Measure for conflicts of this conflict class. The sort key forms an index for the measures within a conflict class. Together with a conflict class key he forms a matrix. With this matrix one can take the appropriate measure for one Look up conflict in the experience base. A conflict class key becomes for example, by the conflict classes of the classification system for Conflicts expressed. The sort key is determined by experience and may change as the experience changes over time. The Experience here is the connection between the circumstances of a Conflict (conflict circumstance) and the success of measures leading to the dissolution to be used by conflicts with this conflict circumstance.

For example, in a railway system can be sorting actions a conflict class depend on multiple sort factors that are in the Assess the success of a measure. Through the operating concept the railway administration can conflict types, measures to resolve Conflicts and a cost function to calculate the cost of a solution To be defined. Another sort factor is the conflict circumstance classes, which are applicable to this very railway system. In particular are this is the abstraction of the location and vehicle parameters, the infrastructure and the Timetables. It can be taken into account which measures in exactly good grip on this railroad system. Furthermore, it can be taken into account how the conflict resolution applies measures, namely incorrect or correct.

The sorting of measures of a conflict class is not constant. The sorting is not determined once and then does not change anymore. Furthermore, the sorting is not deterministic. It can not be theoretical be precalculated since it is based on experience only. The sorting is continues to depend on the sort factors and responds to changes in these Sorting factors. At the start of the conflict resolution system is sorting indefinite, there is no sorting. While at system startup to first construction of the experience base complete solution spaces for conflicts calculated, the sorting becomes progressively statistically more and more relevant. The significance of sorting before, during and after the first Structure of the empirical basis can be quantified with a statistical method become. The sorting applies only to a given traffic system for which they are was created. Once the sort was statistically significant, i. quality Solutions were quickly found, leading to sudden changes in the Sorting factors, such as the cost function, that the Sorting is no longer significant, i. Good solutions will not work anymore found fast. The improvement or expansion of the experience base after changing the sorting factors, it will close again with a delay sorting, which again statistically for the changed sorting factors is significant.

A statistical method can be used to describe the mathematical Correlation between the number of evaluated measures on the the sorting is based and the probability that one is considered bad measure will be successful. This Relationship determines how many statistical events are evaluated must have been so that the sort order is statistically significant. One Example of a statistical method is a discrete random variable with Binomial distribution.

To quickly find a good resolution to resolve a conflict will be Determines how many measures of a conflict class are classified as "good" should. All others are classified as "bad". It is further defined with what probability a measure classified as "bad" for the best solution. If the sorting in the experience base is statistically significant, the solution space is limited by the fact that only used as "good" classified measures for solution calculation become. This limits the number of solutions to be calculated. Within of the restricted solution space becomes the solution with the largest Quality factor, especially with the lowest cost when using a Cost function, considered the preferred solution. It is also possible, only a solution using the best measures for the original ones as well Consequential conflicts to determine.

The sorting of measures within a conflict class can be optimized become. This can happen especially when the few best Measures of a conflict class do not significantly more often provide the best solutions achieve other measures than the same class of conflict. It should be the Classification system can be improved so that the few best Measures of a conflict class much more often to achieve the best solutions as other measures of the same conflict class. The success of Classification system depends on the abstract location and Vehicle parameters. In particular, the parameters and their Value range can be optimized. The value range of the parameters can be extended become. For optimization more than three sorting values like "high", "medium" and "low" for each or some classification parameters introduced become. Furthermore, the mapping of train runs or the infrastructure can the abstract vehicle parameters or abstract location parameters (mapping) optimized later, after the conflict resolution system has been running for some time become. The boundaries between the sorting values "high", "medium", "low" can be changed. In particular, originally be provided that an automatic or semi-automatic mapping of all train runs on sorting values yields approximately the following distribution: "high" = 25%, "medium" = 50%, "low" = 25%. For optimization, the Value ranges are changed so that, for example, then some train runs, which were previously mapped to the value "low", then to the value "medium" are displayed. Furthermore, it is possible to optimize that individual train runs or individual infrastructure sections are later optimized, i. be mapped differently to abstract parameter values than the first automatic or semi-automatic picture. So you can the The spread of disturbances that occur as a result of the application of Measures on such trains occur by having main traffic zones many connections to other train runs or traffic-related bottlenecks or drive over very busy routes.

The scope of the invention also includes a conflict resolution system Implementation of the method according to the invention with a memory for Storing an experience base and means for selecting Measures to resolve a conflict based on experience and resources for determining solutions and the figure of merit of a solution. With a such conflict resolution system can conflict in a track-bound Traffic system resolved faster than with the state of the art Technology.

The scope of the invention also includes a computer program for Implementation of the method according to the invention.

Other features and advantages of the invention will become apparent from the following description of an embodiment of the invention, with reference to the figures of the drawing, the essential to the invention details show, and from the claims. The individual features can be individually for yourself or for several in any combination in a variant of Invention be realized.

drawing

Fig. 1

eine schematische Darstellung des erfindungsgemäßen Verfahrens.

Fig. 2a

einen Lösungsbaum zur Darstellung eines vollständigen Lösungsraums;

Fig. 2b

eine vereinfachte Darstellung einer Matrix zum Auffinden von Maßnahmen einer Erfahrungsbasis;

Fig. 2c

einen Lösungsbaum zur Darstellung eines Lösungsraums bei Anwendung des erfindungsgemäßen Verfahrens;

Fig. 3

eine Matrix zum Auffinden von Maßnahmen in einer Erfahrungsbasis;

Fig. 4a

eine Darstellung der Verwendung eines Erfolgszählers;

Fig. 4b

eine Matrix mit inkrementiertem Erfolgszähler;

Fig. 5

eine Zuordnung der abstrakten Fahrzeug- und Zugparameter;

Fig. 6

eine Zuordnung der abstrakten Ortsparameter; und

Fig. 7

eine Veranschaulichung der Ermittlung einer Konfliktumstandsklasse eines Konflikts.

An embodiment is shown in the schematic drawing and will be explained in the following description. It shows:

Fig. 1

a schematic representation of the method according to the invention.

Fig. 2a

a solution tree for representing a complete solution space;

Fig. 2b

a simplified representation of a matrix for finding measures of an empirical basis;

Fig. 2c

a solution tree for representing a solution space when using the method according to the invention;

Fig. 3

a matrix for finding actions in an experience base;

Fig. 4a

a representation of the use of a success counter;

Fig. 4b

a matrix with incremented success counter;

Fig. 5

an assignment of the abstract vehicle and train parameters;

Fig. 6

an assignment of the abstract location parameters; and

Fig. 7

an illustration of how to determine a conflict circumstance class of a conflict.

In Fig. 1 the inventive method is shown schematically. In an operational concept 1 of a railway administration, valid conflict types 2 are defined from an operational point of view. Example of a type of conflict is train-following conflict on the open track. "Each railway administration has a catalog of conflict resolution measures 3 in its operational concept 1. The assignment for which conflict type which measure is applicable to train 1 and which measure applies to train 2 of a conflict, for example It stipulates that, for each type of conflict, it simply removes those measures that are operationally or technically meaningless in this type of conflict The quality management standard 4 defines how the quality of a solution is measured by the railway administration For example, it is possible to measure the cost in delay minutes weighted by the train category and to add a delay equivalent for lost connections For conflicts 5 , the two components have conflict types and contention classes 6 . Conflict circumstance classes are used to formalize the circumstances of a conflict by means of abstract parameters. The conflict classes of the classification system for conflicts 5 are stored in an empirical basis 7 , with several measures for resolving a conflict being filed for each conflict class. In this case, a matrix allows access to a measure of the empirical basis 7, wherein the matrix contains a conflict class key and sort key, of which the conflict class key points to a conflict class within the empirical basis and the sort key to a measure within a conflict class.

Typically, there is more than one solution to resolve a conflict, namely, at least one solution per action that is applied to the original conflict, and further solutions to all actions that are applied to a consequential conflict. The solution space consists of all solutions to an original conflict, with each solution being assigned a quality factor. These solutions are obtained by applying to the original conflict and to all subsequent conflicts all applicable measures (according to the type of conflict). First, all applicable measures are applied to the original conflict. Each measure results in a new situation, usually with consequential conflicts. Then, in each of these new post-conflict situations, the first temporal conflict is resolved by applying all the measures applicable to that type of conflict. This in turn creates a new situation per applied measure. This will continue until there are no follow-up situations with conflicts or until conflicts are so far in the future that they are no longer interested. For example, when n, each solution to an original conflict of uniformly = 4 measures is uniform and the same for each original or subsequent conflict m number of applicable steps has, is the solution space of m ⁿ solutions. For n = 4 and m = 8 there are 4096 solutions. If instead of calculating all solutions for all measures, applying only the best two measures of a conflict class, this narrows the solution space in the above example to only 16 solutions. Therefore, the figure of merit has to be determined for much fewer solutions.

This will be explained again with reference to FIGS . 2a to 2c . In the case of the solution tree 10 shown in FIG. 2a, the situation S0 represents an original conflict case. In order to resolve this conflict, it is possible to use one of the measures M1, M2, M3 . For each applied measure, this leads to the follow-up situations S1, S2, S3, which each represent a consequential conflict, to which in turn various measures can be applied in order to resolve it. For the situation S1, these are the measures M11, M12, which lead to the follow-up situations S11, S12 . The consequential conflict S2 can be resolved by the measures M21, M22, M23 , with situations S21, S22, S23 occurring in each case. Accordingly, the measures M31, M32, M33 can be applied to the situation S3, which leads to the follow-up situations S31, S32, S33 . In the example of FIG. 2a, situations S11, S12, S21, S22, S23, S31, S32, S33 do not represent conflict situations that need to be resolved. Each path (path) through the solution tree 10 from the original conflict S0 to a conflict-free situation S11, S12, S21, S22, S23, S31, S32, S33 represents a solution. A complete solution tree in which the original conflict and each contention conflict all applicable measures corresponds to the solution space. In the prior art, a quality factor for the solution is determined for each solution. In order to obtain the best solution, ie the one with the best quality factor by comparison, the entire solution space must be determined. In the example, the solution 11 is the best solution.

In FIG. 2b, by way of example, a matrix 12 is shown simplified. Each situation or conflict S0, S1, S2, S3 is assigned to a conflict class KK1 - KK4 . In particular, the conflicts S0, S1, S2, S3 are assigned to the conflict classes KK2, KK1, KK4 and KK3. The applicable measures are stored for each conflict class KK1 - KK4, the measures being sorted according to their chances of success. The same measure can exist in different conflict classes. For example, M1 and M23 may be identical. For the conflict category KK2, measure M1 is the most promising measure and M3 the least promising measure. If in the conflict resolution system it is given that only the two measures for the resolution of a conflict are to be used to find the best solution, which promise the most success, one obtains the (incomplete) solution tree of Fig. 2c.

On the original conflict situation S0 only the measures M1, M2, the the follow-up situations S1, S2 lead, applied. On the follow-up conflict situations S1, S2 are respectively the measures M11, M12 or M22, M21 so that the conflict-free situations S11, S12, S22, S21 are obtained become. The quality factor must therefore only for the solutions to the Situations S11, S12, S22, S21 lead to be determined. By comparing the Quality factors in turn gives the solution 11 as the most favorable solution, but much faster, because fewer solutions need to be calculated.

FIG. 3 shows a matrix 20 for finding measures in an empirical basis. The conflict class key K, which is shown on the abscissa, together with the sort key SS (shown on the ordinate) for the measures M _a - M _i within a conflict class KK1 - KK7 forms a matrix 20. The applicable measures M _a - M _i for each conflict class KK1 - KK7 sorted according to their success under the conflict circumstances of the respective conflict class KK1 - KK7. With this matrix 20, the applicable measure M _a - M _i can be looked up for a conflict in the empirical basis. A measure M _x may be applicable in more than one conflict class. The measure M _g is applicable, for example, to the conflict classes KK1, KK3, KK4 and KK6. For the conflict class KK1 it is the best measure, for the conflict class KK3 it is the eight best measure, for the conflict class KK4 it is the third best measure, for the conflict class KK6 it is the fifth best measure.

For each conflict class, there is one success counter, not shown in FIG. 3, for each applicable measure. For measure M _g, there are four separate success counters: one each for conflict classes KK1, KK3, KK4 and KK6.

The basis of experience is built up by the evaluation of completely calculated solution trees. If a conflict of the conflict class KK1 has best been solved by the measure M _g , then the success counter for M _g in the conflict class KK1 is increased by 1. If a conflict of the conflict class KK3 is best solved by measure M _g , then the success counter for M _g in the conflict class KK3 is increased by 1, etc. After each change of a success counter, the sorting of the measures within the corresponding conflict class is checked and if necessary tightened.

The use of the success counter Z _x will be explained again below with reference to FIG. 4 a. The success of a measure M _x is the frequency with which this measure M _{x has led} to the best solution (lowest cost) of conflicts under certain conflict circumstances (conflict class). To be able to determine this frequency, success counters Z _x are used. For each conflict class CC1 - KK7 there are per applicable measure M _x a success counter Z _x. A measure M _x can be applicable in more than one conflict class KKx. Then there are for this measure M _x several success counter _x Z. For each applicable measure in the above sense, there is a success counter (per conflict class), ie within the same conflict class you need for "hold extended at turn 1" and "hold extend at turn 2" two success counters. In FIG. 4a, the measure M _{e is} applicable in the four conflict classes KK2, KK4, KK5 and KK7. Therefore, for M _e, there are four separate success counters Z _e : one each for the conflict classes KK2, KK4, KK5 and KK7. In FIG. 4 a, all success counters Z _x are initially set to the value 100.

When the experience base is built, the conflict class is determined to conflict K from the solution tree, for example KK6. Furthermore, it is determined which measure applied to conflict K led to the best solution, for example measure M _c . The success counter Z _c for measure M _c for conflict class KK6 is determined and incremented ( FIG. 4b ). After each change of a success counter Z _x , the sorting of the measures M _x within the corresponding conflict class KKx is checked in the empirical basis and, if necessary, followed. Variants for counting success are conceivable. The counting can for example be combined with a weighting factor. Other variants relate to which and how many measures of a solution tree are taken into account in the success counting. The evaluation can be restricted to only the measures applied to the original conflict (ie only M1 in Fig. 2c) or to the measures applied to all conflicts of the best solution (M1, M12 in Fig. 2c), or consider any measures applied to any resolved conflict of the solution tree that have best resolved this (consequence) conflict (M1, M12, either M22 or M21, depending on which of the two corresponding solutions is more cost effective).

FIG. 5 shows the assignment of the abstract vehicle or train parameters. Depending on a timetable FP , eg a winter or summer timetable, a mapping to abstract train parameters is made for each train number NR . The train with the train number NR = 4711 is a train priority ZP with the value "low" N , a train power ZL with the value "medium" M , a train run priority ZLP with the value "medium" M and a priority of line recovery PLW with the value assigned "high" H Accordingly, the train with the number NR = 0815 the parameters ZP = H, ZL = N, ZLP = M and PLW = H assigned.

FIG. 6 illustrates the assignment of the abstract location parameters. For every possible conflict location, the infrastructure IS maps to abstract location parameters. The location with the location identification number Ol = 1234 is assigned a diversion potential UP with the value "low" N in both directions R. The location with the number OI = 1235 is assigned in one direction R the bypass potential UP = H (high) and in the other direction R the bypass potential UP = N.

FIG. 7 shows the determination of the contention class KU of a conflict between the trains with the number NR = 4711 (train 1 Z1 ) and NR = 0815 (train 2 Z2 ). From the abstract train and location parameters in the upper half of Fig. 7, ie above the line L , conflict formalization parameters are determined which have the same designation as the abstract parameters. The conflict formalization parameter train priority ZP receives the comparative value Z1 <Z2, the train power ZL the value Z1> Z2, the train run priority ZLP the value Z1 = Z2, the priority of the line recovery PLW the value Z1 = Z2 and the diversion potential the value Z1 = Z2. All conflicts with the same combination of values for the conflict formalization parameters belong to the same conflict category KU. In the case of a conflict with only one train, it can be provided that this train 1 Z1 is compared with a reference train Z2 which has the value "medium" in all abstract parameters.

In a method for resolving conflict situations (S0, S1, S2) in a track-bound transport system will be a selection of measures (M1, M2, M11, M12, M21, M22) suitable for resolving the conflict, taken from an empirical basis and only these measures (M1, M2, M11, M12, M21, M22) leading to follow-up situations in the determination a solution (11) considered. For each solution a quality factor is determined and then, by comparing the quality factors, the best solution (11) determined and applied.

Claims (12)

1. Method for controlling the occupancy in a trackbound transportation system in the event of an occupancy conflict (S0, S1, 52, S3) or a technical conflict, possible alternative measures (M1 - M33; M_a - M_i) being determined for the resolution of the conflict (S0, S1, S2, S3) and being taken into consideration in the determination of a solution, a merit factor being determined for each solution, and the solution (11) with the best merit factor being applied, characterized in that a selection, comprising not all of the measures possible for the resolution of a conflict, of measures (M1, M2, M11, M12, M21, M22) possible for resolving the conflict are taken from an empirical basis (7), the success of a measure (M1 - M33, M_a - M_i) being evaluated and only the measure(s) classed as promising (M1, M2, M11, M12, M21, M22) being taken from the empirical basis (7) and only this (these) measure(s) (M1, M2, M11, M12, M21, M22) being taken into consideration when determining solutions (11).
2. Method according to claim 1, characterized in that for a conflict (S0) a selection of measures (M1, M2) is taken from the empirical basis (7) and for each resultant conflict (S1, 52) arising due to a measure (M1, M2) applied to the conflict (S0) a selection of measures (M11, M12, M21, M22) is likewise taken from the empirical basis (7).
3. Method according to claim 1, characterized in that a sorting key (SS) is defined, via which the suitable measures (M1 - M33) for a conflict to be resolved can be found in the empirical basis (7).
4. Method for generating an empirical basis (7, 12) characterized in that the empirical basis (7, 12) is built up in operationally slack times of a conflict resolution system, in which the method according to claim 1 is implemented, and/or following a first implementation of the method according to claim 1, in particular the initial installation in the conflict resolution system.
5. Method according to claim 4, characterized in that conflict types (2) and conflict circumstance classes (6, KU) are defined, which are assigned to a conflict class (5, KK1 - KK7) and the conflict classes (5, KK1 - KK7) are filed in the empirical basis (7).
6. Method according to claim 5, characterized in that the applicable methods (M1 - M33, M_a - M_i) for the resolution of a conflict are assigned to each conflict class (5, KK1 - KK7).
7. Method according to claim 6, characterized in that the applicable measures (M1 - M33, M_a - M_i) within a conflict class (5, KK1 - KK7) are sorted.
8. Method according to claim 5, characterized in that conflicts are classified with reference to their formalized circumstances by the conflict circumstance class (6, KU).
9. Method according to claim 8, characterized in that the classification is achieved with reference to abstract location and vehicle parameters (ZP, ZL, ZLP, PLW, R, UP).
10. Method according to claim 6, characterized in that a success factor (success counter Z_a - Z_i) is assigned to each applicable measure (M1 - M33, M_a - M_i) of a conflict class (KK1 - KK7).
11. Conflict resolution system for executing the method according to claim 1, with a storage device for storing an empirical basis and with means for selecting measures for the resolution of a conflict from the empirical basis and means for determining solutions and the merit factor of a solution.
12. Computer program for executing the method according to claim 1.

Images