DE19963590B4 - Method for controlling transport units in a traffic network - Google Patents

Abstract

Method for controlling transport units in a traffic network,
Having interconnected path elements occupied by a maximum number of transport units predetermined for each path element, the predetermined number being variable during the course of the method,
The time is divided into discrete time intervals and each transport unit begins its journey at any time interval,
Each path element is covered by a minimum charge to be paid by a transport unit for the time interval at which it occupies the path element,
- each transport unit is provided with a budget of charging units, and
- The transport units bid for a way-out using an auction mechanism.

Description

The The invention relates to a method for controlling transport units in a traffic network that has interconnected path elements, with a for each path element predetermined maximum number of transport units be occupied, the predetermined number during the course of the process is variable.

Of the The state of the art distinguishes between scheduled and timetable-free transport systems.

One typical example of a scheduled transport system is a railway network. Currently becomes the control of trains in rail networks through central units (namely the signal boxes), that adhere to a global roadmap, adherence to this roadmap monitor all trains and make corrections in case of deviations.

These corrections are still largely done in practice by human intervention. There are also already theoretical approaches to automatically adapt the timetables in real operation when deviations from target behavior are detected. Examples of such approaches are described in [H. Dannenberg: Ordered Search Algorithms for the Computer-Aided Disposition of Railway Operations, Series Scientific Works No. 23, Institute for Transportation, Railway Construction and Operation of the University of Hanover, 1986] and [ DE 197 26 542 A1 ]. However, all known approaches continue to assume that the trains operate at all according to timetables, which are managed centrally.

One typical example of A timetable-free transport system is a road network. The vehicles are controlled completely autonomously and are not subject to any central Control. But it is useful, automatically recommended routes for the vehicles to give consideration to the individual wishes of the driver wear. First versions of such route recommendation facilities are already commercially available in production vehicles. These are neither on individual preferences the driver still on the current traffic situation (for example congestion). Approaches, who take that into account still in the research stage.

From the DE 44 16 125 A1 For example, a method is known for controlling transport units in a traffic network having interconnected path elements, wherein the time is divided into discrete time intervals and each transport unit begins its journey at any time interval, and each path element is allocated a minimum charge, which is one Transport unit is to pay for the time interval to which it occupies the path element.

Out WO 00 542 40 A1 likewise discloses a method for controlling of transport units in a transport network that are interconnected Way element, each transport unit to any Time begins its journey, and each path element is charged a fee, that of a transport unit for the time interval is to be paid, to which it occupies the path element.

Kumar, J .; Sheble, G .: Option Market Simulator For Price Based Operation, in IEEE Transactions on Powersystems, Volume 13, no. 1, February 1998, p. 250 - 255, discloses a method for controlling the power supply in one electric power grid, with transmission capacities with Help auction auction mechanism.

Of the Invention is based on the object, a method of control of transport units in a transport network with an optimized Specify operating procedure, which is used for traffic systems a timetable is unnecessary and takes into account the current traffic situation.

The Task is solved by a method for controlling transport units in one Traffic network having interconnected road elements that with a for each path element predetermined maximum number of transport units be occupied, the predetermined number during the course of the process is variable, according to the invention the time is divided into discrete time intervals and each transport unit at any time interval can begin its journey, each Path element occupied by a minimum fee is paid by a transport unit for the time interval is to which it occupies the path element, each transport unit Budget of fee units to disposal is placed, and the transport units are a Wegbenutzung bid using an auction mechanism.

The motivation for the present invention is to improve the operation in traffic networks. The invention can be used for transport networks such as rail traffic, air traffic, shipping traffic and individual road traffic and offers advantages over the prior art for these transport networks. The invention can be used wherever "transport units" are moved from one location to another should. Railways, road vehicles, ships, aircraft, but also containers or, for example, parcels are thus conceivable as "transport units." For the sake of clarity, the other exemplary embodiments are generally directed to rail traffic, but are also applicable to all other conceivable embodiments of transport nets applicable.

in the Contrary to the prior art, the method of the present invention Invention completely without timetables out. Trains For example, you drive individually through your rail network and you are through controlled by their own decentralized software units. It is advantageous but not necessary, these decentralized software units as software agents to implement. The decentralized approach makes it possible to be more flexible to the unforeseen disorders to respond, and only make the recalculations that really what is needed, which significantly reduces complexity and thus the ability decisively increased, to use the controller during operation. The operational This process is optimized.

By the waiver of central timetables will Furthermore offset a competitive disadvantage of the railway opposite the road: The fact that a road vehicle anywhere at anytime can make the way and is not centrally controlled and on this Way react more flexibly to disturbances can, causes many people prefer to use a car than the train, and many Freight forwarders, their goods prefers to be transported by truck rather than by freight train. The present invention allows it's the trains, similar in the rail network free to move like a car in the road network.

The Assignment of the different routes takes place via an auction process, where the trains are participate with a given budget. An auction process has the advantage that it free to the auction participants (in this case the trains) what criteria they choose for a cheap route, and still takes care of "soft Kind "for that it too a just balance of conflicting interests comes because that Budget for each participant is limited. Through different sized budgets can certain auction participants are prioritized without them being weaker applicants automatically beat out of the field (a less fortunate auction participant, the only a short distance needed for the Others are interested, and no substitutes has, may be offer more than a "rich" auction participant, which can drive on a different route at low additional costs). In this way, the process meets the requirements of rail transport fair, between trains different categories such as ICE, D-train and regional traffic.

There the present method is not dependent on the vehicles track-bound, it can also be used in air traffic. All just mentioned advantages also apply to this traffic area, because aircraft do not fly on the traffic routes Free around, but stick to flight plans and will be right now controlled by central flight control centers. Offers a special advantage the present proceedings in the catchment area of airports, in which the air traffic is very overloaded is and often leads to "traffic jams", what in waiting for the individual aircraft affects.

The present method can also be used in individual road traffic. In this traffic network, even in the prior art, the vehicles can move without central control, so that the method of the invention does not mean a substantial improvement in this regard (unlike the above-mentioned traffic networks). However, there are characteristics of the method which mean an improvement not only for track-bound means of transport and for aviation, but also for individual road traffic:
The method makes it possible for each vehicle that is on an already pre-planned route, in a spontaneously occurring event (usually a "fault") either for the route or for the vehicle to redesign the route. B. traffic jams are avoided in traffic.

For the mere bypass of an existing congestion, the method used offers the advantage to the state of the art that it individual preferences when recommending replacement routes and objectives of the user There are, for example, not all affected vehicles the same recommendation.

The Auction procedure provides additionally for one even distribution the vehicles on the roads, without all vehicles being treated equally become. In this way, the process not only serves the congestion avoidance, but also the congestion reduction.

Further advantageous embodiments are given in the dependent claims.

The The invention is based on a FIG, which is a driving-waiting search tree in a schematic illustration, and several embodiments explained in more detail.

The method of the invention uses the following model for the traffic network:
The smallest units of the traffic network are path elements. These are interconnected. The locations where one way element may be connected to another is commonly called a port. Path elements are usually assigned to type classes. Each type class has well-defined ports and parameters that determine how to get from one port to another. In this way one can define way elements of the type switch, track element, crossing, etc. The example of Weiche makes it clear that it is not always possible to get from each port of a path element to every other port of the same path element. The parameters of a path element also specify the length, geographic location, and other physical conditions that determine how quickly a transport unit can travel from one port to another. Stations or other breakpoints must also be defined as a way element. This corresponds to reality, because even breakpoints have in reality a spatial extension.

The just listed Terms are mostly taken from the railways. In particular, not track bound Transport networks are usually modeled differently: In general Case consists of a traffic network of nodes and connections between this node. The control task consists of transport units from one node of the transport network to another. Of the Transfer to the model described above is easy: nodes and connections are each modeled as a path element or as a sequence of path elements. The control task then consists of transport units of to guide one path element to another.

Several Transport units are allowed use the same way element, but in railway traffic not at the same time. It is common in this domain with discrete time intervals to work, to which a train is a way element can prove. A path element may only be in the same time interval be used by a transport unit. The occupancy of an entire Way over several time intervals is called the schedule route.

Also the present method operates at discrete time intervals. Since the procedure is not limited to the railway traffic and does not presuppose the existence of a timetable is the term Timetable route misleading and will be replaced in this description by the term travel use. The control task is therefore one for each transport unit Find way use, i. the determination of consecutive time intervals for the Elements of a path, so that it there are no conflicts. A conflict occurs when more transport units as admitted at the same time prove a way element.

As just mentioned, is the allowable maximum number of transport units per time interval in railway traffic at 1. The same is true for the Air traffic makes sense. If you have the same principle in road traffic you would have to apply for a good one Utilization of physically and safety-permissible density the vehicles both the time intervals and the path elements infinitesimally small, resulting in an extremely complex model of the transport network would. For this reason, the procedure leaves to that way elements also occupied by several transport units in the same time interval be allowed to i.e. the specified maximum number of transport units may greater than Be 1. This maximum number is always a clearly assigned parameter, which depend on the type of path element may and possibly other parameters (e.g., length). To the Example, you can use a three-lane highway route in Wegelemente the length 10 km split and a maximum occupancy set of 600 vehicles per minute, with an average of Transport units would occupy the path element for 6 minutes. The Maximum number of occupants may fluctuate in the course of the control procedure: for example, the Blocking a track the maximum number of seats of the path element just described to 400 vehicles.

each Path element is occupied by a minimum fee, that of a transport unit is paid per used time interval. This is it irrelevant whether the transport unit moves on the path element or if it just waits there, if standing still for this Path element allowed is. Each transport unit receives a budget of charging units and buys in an auction that works for all path elements of the network is held, a Wegbenutzung, i. the occupancy of the elements a path at certain times.

each Transportation unit can in principle get a different budget as well as a different strategy, as this budget uses shall be. An advantageous embodiment of the method shares the vehicles in different categories, for each of which a generic Budget and a generic strategy is set for all vehicles of the category concerned. Of course it is possible that individual Vehicles continue to receive individual budgets or strategies even if they belong to a category, in the sense of an object-oriented one Inheritance hierarchy.

The vorteilhaf described below Embodiments are in principle applicable to all transport networks. But they are particularly suitable for their use in transport networks, in which the maximum occupancy number for each path element per time slice must be one, for example in rail transport. Later in the description, embodiments are described that are particularly suitable for use in transport networks, in which the maximum occupancy number for each path element per time slice is much greater, eg in traffic.

The following Design of the method is the basis for all further embodiments, that respect the special needs of rail transport take: A data processing unit acts as an auctioneer. Every transport unit that wants to bid for a way-out, must be at Sign up for auctioneer. The auctioneer leaves any interested transport unit an offer on a list of the required routes with the associated time intervals submit. An auction mechanism determines which offer as the best is classified and in which order the transport units be questioned. For example, there may be auction rounds in which In turn, each transport unit may submit exactly one offer, the for each single path element outstrips all offers available so far. The minimum bet must be in in any event, be the minimum fee set in the procedure. Each transport unit, which receives the surcharge for a Wegbenutzung has, is the last for it offered fee deducted from his budget.

• 1. Fahrwegsuche: Bestimmung eines neuen Weges w vom Ausgangs-Wegelement zum Ziel-Wegelement
• 2. Errechnung einer Wegbenutzung zu vorgegebenem Weg w: Bestimmung der benötigten Zeitintervalle für die Wegelemente des Weges w, wobei auch Wartezeiten zulässig sind.
• 3. Berechnung eines Angebots zur Abgabe an den Auktionator: Für bereits in Schritt 2 bestimmte Wegbenutzungen wird ein Angebot berechnet, das für jedes betroffene Wegelement die Mindestgebühr und das bisherige Höchstgebot übertrifft, unter der Maßgabe, daß das für die Transporteinheit festgelegte Budget nicht überschritten wird.

For each transport unit, there is a local data processing unit that solves the following tasks:

• 1st route search: determination of a new route w from the exit route element to the destination route element
• 2. Calculation of a distance to a predetermined path w: Determination of the required time intervals for the path elements of the path w, with waiting times are also allowed.
• 3. Calculation of an offer to be submitted to the auctioneer: For uses already determined in step 2, an offer is calculated which exceeds the minimum fee and the previous maximum bid for each element concerned, provided that the budget set for the unit of transport is not exceeded ,

These Tasks can different times and edited in any order, and moreover be interwoven with the various auction rounds. advantageous Embodiments as may be more specifically described will be described later.

given Let be a calculated in step 1 track w. In a further embodiment Now an auxiliary graph is constructed to use a search algorithm for the shortest routes the cost-optimal solution to get.

In In the following description, the search algorithm will be the algorithm used by Dijkstra. The advantageous embodiment works but also with every other way algorithm. For the calculation of your own Offer are in the following the present highest bids for the respective Path unit and the respective time interval the basis. The commandments, If necessary, the minimum price of a time interval, become the cost for the fastest possible way use summed up.

The Decision, the fastest possible way of use next Select offer sets the readiness of the transport units ahead, all already existing Bids on the required time intervals to outbid. In addition to "outbidding" there is the possibility on later To wait for time intervals that may be cheaper, with the "wait" by the whereabouts on the old way element also causes costs.

By the possibility, In each time interval, it is possible to opt for "wait" or "outbid", as is the case with many Search problems multiple valid Solutions. It will be for the predetermined path w sought the offer, the minimum cost caused.

Of the to be constructed driving-waiting search tree is schematic in the FIG and has as nodes pairs of the form (path element of w, set of time intervals). The goal is, starting from the root node A = (Initial path element, initial time interval) to a node B. find whose path element is equal to the destination path element.

each Edge of the search tree corresponds to a decision on whether the existing bid for a lot of needed Time intervals outbid should or should be on a later Time interval is to be maintained. The travel costs of each edge are the lowest bid, all the others at the present time Offers already made in the auction for the corresponding path element and time interval exceeds.

A restriction of the solution space is possible by limiting the maximum waiting time. Calculated this will be the cost of the fastest possible SWB use. SWB is the path P (A, B), which results when the transport unit in any case for to outbid of the already existing offer decides.

at the further search only those paths are to be considered their cost is less than the cost of the fastest possible Road use SWB are.

• – costs[i] enthält die Kasten des kostenoptimalen Weges von A nach i.
• – predecessor[i] enthält den direkten Vorgänger von i auf dem kostenoptimalen Weg von A nach i.
• – OPEN ist leer.
• – CLOSED enthält alle Knoten.

The following describes how Dijkstra's algorithm solves the path problem:
The algorithm works with the fields costs [1..n], predecessor [1..n], and the sets OPEN and CLOSED. These objects should contain the following data at the end:

• - costs [i] contains the boxes of the cost-optimal route from A to i.
• - predecessor [i] contains the direct predecessor of i in the cost-optimal way from A to i.
• - OPEN is empty.
• - CLOSED contains all nodes.

• – costs[i] enthält die Kosten des kostenoptimalen bereits errechneten Weges von A nach i.
• – predecessor[i] enthält den direkten Vorgänger von i auf dem kostenoptimalen bereits errechneten Weg von A nach i.
• – CLOSED enthält alle Knoten, von denen bereits Kinderknoten erzeugt wurden
• – OPEN enthält alle noch aktiven Knoten (alle die, von denen noch keine Kinderknoten erzeugt wurden).
• – Ein Iterationsschritt verläuft nun folgendermaßen:
• – Es wird der Knoten k aus OPEN ausgewählt, für den costs[k] minimal ist.
• – Wenn k die Abbruchbedingung eines oben beschriebenen Knoten B erfüllt, dann ist der Algorithmus beendet.
• – Wenn k die Abbruchbedingung nicht erfüllt, dann bilde zwei neue Kinderknoten nach der Konstruktionsregel des oben beschriebenen Suchbaums.
• – Verschiebe k nach CLOSED

In the iteration, successive nodes are shifted from OPEN to CLOSED. After each iteration, the objects should contain the following data:

• - costs [i] contains the costs of the cost-optimal already calculated route from A to i.
• - predecessor [i] contains the direct predecessor of i on the cost-optimal already calculated path from A to i.
• - CLOSED contains all nodes from which child nodes have already been created
• - OPEN contains all nodes that are still active (all those from which no child nodes have yet been created).
• - An iteration step now proceeds as follows:
• - The node k is selected from OPEN, for which costs [k] is minimal.
• If k satisfies the abort condition of a node B described above, then the algorithm is finished.
• If k does not satisfy the termination condition, then create two new child nodes according to the design rule of the search tree described above.
• - Move k to CLOSED

In another embodiment of the method, the case is considered, that the Transport unit at a proposed occupancy of a path element via a certain number of time intervals of this path element with a must pass through certain speed. Because the transport unit In this way it receives a minimum braking distance, it is possible that it too when using the maximum braking force is no longer on the same Wegelement can come to a standstill. So it can be on this Do not wait any longer for the waypoint. The advantageous embodiment marked the search tree node with the maximum possible waiting time and forms no successor nodes that use the same path element assignment require a time interval after this maximum possible waiting time.

The Procedure is basically terminated by the fact that the Auctioneer no more auction rounds more permits. Then no transport unit can bid more new ways to use and has to be satisfied with the previously purchased.

Various Embodiments of the method can determine when the auctioneer cancels the auction. One obvious abort criterion is that all transport units are have auctioned off one way use, no way at the same time interval over his maximum number is used and each transport unit with their use is "satisfied", i.e. none Transport unit more need for sign up for another way-use.

One of the significant advantages of the method of the invention is that the transport units can react very flexibly to unforeseen disturbances. An embodiment of the method therefore describes how it is recognized that a transport unit can not realize an intended path utilization and how such a transport unit can then obtain a new path usage:
The reason why it is no longer feasible to use a route can be due to unforeseen changes along the way (such as blockade of the line, traffic jam) or due to the transport unit itself (eg reduced maximum speed, delay due to higher passenger numbers) , Therefore, not only the transport units but also the path elements are assigned local data processing units which monitor any changes in the parameters. If a fault is detected on a path element, then all transport units are informed that want to use the path element and could be affected by the change of the parameters th. If a transport unit by such a message or by a fault caused by himself determines that there is a It can no longer realize the use of a route, it tries to bid for a new way use. For this, it must either trigger a new auction (as at the beginning of the procedure, if there is no assignment of wayside use) or participate in an auction triggered by another transport unit.

At the beginning of a new auction, new budgets are distributed to the transport units. Otherwise, a transport unit that has already used up its entire budget for the previously known conditions could no longer react to unforeseen changes. Nevertheless, will be Do not involve all transport units in a new auction: Transport units whose off-road use are not affected by the rescheduling of other transport units retain their way-off and continue their journey as planned.

A Strategy has the task of deciding when and under what Condition that one of the actions described above be executed should.

A Possibility, defining a strategy is between the two actions to decide if a new route should be calculated or whether a new way use for an already calculated path and a corresponding offer are calculated should. Another possibility is to decide whether to calculate a new road use should be or whether an already calculated use of a road new offer should be calculated.

In a further embodiment are four train categories, ie categories for transport units, for rail traffic set: ICE, IC, IR, RE and GV.

at Any bidding will try an ICE first, its originally planned Off-road use (probably the fastest way to the destination) consistently outbidding all competitors for to prove themselves. Only when the cost of this way 1.5 times exceeds the cost of using the same route in a later time interval, the ICE gives way to the later Time interval off. Exceed the price for a way-use in a later Time interval 2 times the current cost of another way, gives way the ICE to use this route.

at Any bidding will try an IC first, its originally planned Off-road use (probably the fastest way to the destination) consistently outbidding all competitors for to prove themselves. Only if the costs for this use 1.3 times exceeds the cost of using the same route in a later time interval, the IC deviates to the later time interval out. Exceed the price for a way-use in a later Time interval 1.5 times the current cost of one the other way, the IC deviates from using this driveway.

One IR compares the costs by outbidding for the originally planned road use with the costs of using the same infrastructure in a later one Time interval and occupies the cheaper variant with a bid. Exceed the cost of both variants 1.3 times the cost of use another way, the IR dodges the other way.

One RE compares the costs by outbidding for the originally planned road use with the costs of using the same route in a later one Time interval and occupies the cheaper variant with a bid. Exceed the cost of both variants 1.3 times the cost of use another way, the RE deviates the other way.

One GV compares the costs by outbidding for the originally planned road use with the costs of using the same route in a later one Time interval and the cost of the Use a different route and choose the cheapest option.

Independent of In their strategies, the train categories are still with different Budgets equipped: So receives an ICE 10 times, an IC 5 times, an IR 3 times and a RE is 1.8 times the budget of a GV.

A another variant is in the presence of a timetable special delay costs and to define these in the just mentioned cost comparisons to take into account.

For every category of transport units can thus factors n and m (n <m) To be defined. The transport units try to calculate a new way and one for it intended use of the route as long as exclusively this way use by overbidding to bid on all competitors until the fee to be paid is n-fold the first calculated fee is. First subsequently is resorted to another way use the same way, the for a later Time interval begins. The transport units only then pull one considering the new route if the fee to be paid for the other Distance of use is m times the fee calculated for the first time.

Thus, the description of the advantageous embodiments of the method is completed with particular regard to the rail traffic. The following aspects take into account the specific concerns of road traffic:
Unlike rail vehicles, it is not common for road vehicles to stall for coordination with other road users and wait for the next line usage license. A method requiring this would not be accepted by road users. Therefore, it does not make sense in road traffic to charge the fee for each time segment, but only for the use of a particular We regardless of how long it lasts. In this way, a vehicle that is in a traffic jam, not punished additionally.

It It is also reasonable to require that the maximum occupancy number for a Path element not only from the current traffic situation, but also depends on the required transit time. This attribute is in traffic Fulfills: If a piece of road at a certain number of vehicles, ie. all Vehicles drive in column a certain speed, then would the Addition of other vehicles the possible speed of all Reduce vehicles. But if you have a longer transit time and thus a lower speed than the currently possible a priori, then you can get more vehicles on the street piece at the same time Allow time.

in the road traffic the auction process should be slightly different than the one described above Configurations: Because of the large number The road user is not practicable if every single one to trigger an auction may. Rather, this task should be undertaken by a vehicle coordinator Be the auction wishes and decides after its own heuristic whether one triggered new auction shall be. The calculation of an optimal temporal use must be in road traffic not so much in the foreground as in the rail traffic, why the bidders in their bid are held only one of them as the best valued way and not the temporal use to book it exactly.

When meaningful addition to this process could the vehicles their estimated temporal use of the path elements calculate and submit in the offer. That way the auctioneer could Assign more vehicles to the same route than without the knowledge the usage times. This procedure is also without one during the Driving continuous time monitoring and enforcement the reservations on average effectively. The goodness depends essentially of kindness the prognosis of when the vehicle is expected which path element will use.

in the road traffic is not it realistic that one Method influence on all moving on the street Vehicles has. Therefore, proceed an embodiment of the method in the allocation of vehicles so that it the vehicles on the route or predicted to which it has no access, deducts from the maximum occupancy and only the difference is distributed.

Further Embodiments specify details of the process designed for road traffic. In the following a sequence of the procedure is reproduced, which all contains these details.

The triggering of an auction happens as follows:
The heuristic of the vehicle coordinator triggers an auction after a sufficient number of vehicles have registered an auction request: these are either vehicles that are to be provided with a route for the first time so that they can start their journey, as well as vehicles already in motion whose route has deteriorated significantly. This list will now be supplemented by the vehicle coordinator to include other vehicles: All vehicles to which the route could be challenged by the aforementioned vehicles must also participate in the auction. To determine are thus all vehicles whose route contains at least one route segment to which another vehicle makes a new bid and in which the requested time slices of the other vehicle overlap with the auctioned time slices of the first vehicle. Thus, for a given route segment and a predetermined set of time slices, it is necessary to determine all the vehicles which have acquired the use of the route segment for one of these time slices. In order to keep the circle of potential auction participants smaller, one should limit oneself in the consideration of the time slices to a limited number into the future. Conflicts for long-term time slices will be resolved in later auctions. Under certain circumstances, these conflicts no longer exist because of new changes in the transport network.

In every new auction receives each participant in this auction has a new budget. Not yet used up Budget from past auctions is available to every auction participant additionally to disposal. It is thus up to the individual bidding strategy, get some budget for to lift unexpected obstacles in the future. In addition, everyone receives Auctioneer a refund for already purchased and thus "paid" route parts, which he then did not drive. In this way, vehicles that have in previous auctions have come too short or whose auctioned Route proved in the following as not passable, in subsequent Auctions automatically get more opportunities because of their budget through non-utilization from driveways grows.

Each auction consists of several rounds of auctions. At the beginning of each auction round, each vehicle identified as an auction participant according to the criteria described above sends the message to the vehicle coordinator it is interested in carrying out a new auction round at all. A heuristic of the vehicle coordinator decides whether a new auction round will be carried out. This heuristic takes into account the size of the share of auction-willing vehicles in the totality of all potential auction participants.

If a new auction round is being carried out, everyone has to Auction participants (even those who did not originally have a new auction round wanted) bids on all segments of a selected route submit. For each segment becomes the desired one Usage period specified. This period of use is calculated itself from the data supplied with each segment and the starting time of the vehicle: This will be the time of the expected entry and exit calculated from the segment. The requested period of use should cover these times generously, since it is only a forecast, which is given under circumstances can still change. This is especially true for Segments that are still far from the current location of the vehicle are removed. Every bid must be at least be as high as the minimum fee required by the segment. These fee refers to the use of the segment from beginning to end and does not hang of the length of the requested period of use.

Everyone Auction participant makes a bid on exactly one route. It is It is up to the bidding strategy of the vehicle to decide which route it is among the previously calculated ones and what proportion of its budget it uses on the route it chooses, furthermore, how is the total budget for the route to the individual Distributed segments. It remains subject to the bidding strategy of the Vehicle, how generous the requested period of use for a segment.

Of the Vehicle Coordinator collects the bids and directs them closed to the auctioneer, i. the submission of offers in one Auction round runs parallel. Therefore, vehicles do not have the option in their bidding behavior the bids of the competitors in the same Auction round to consider. The vehicles have the option to view the bids of the last auction round anonymously, i.e. they learn how many bidders are bidding for each Have submitted segments. But they do not know the identity of the bidders and can Therefore, do not learn the bidding behavior of the competitors.

Of the Auctioneer now redistributes the requested segments a specified auction mechanism to the bidders. following Auction mechanism is for the method particularly advantageous: All auction participants offer parallel. The contract will be awarded to the highest bidder, but at the price of the next bidder. This auction mechanism is the first-bid-secondprice in the literature known. He is not very time consuming and particularly advantageous for honest bidders, i.e. It is not worthwhile to offer less than the bid is really worth it. For the process of the invention must be Auction mechanism be slightly modified because in the considered Scenario for multiple bidders who are assigned the same resource can (because the maximum occupancy of a segment is much larger than 1): The price of the highest Bidding of a vehicle that has not got the route anymore, applies to all who got the track. If all bidders the route have received, the price is the minimum fee.

Of the Auctioneer takes the wishes the vehicles and checked for each Segment, how many vehicles it depends on current occupancy and maximum occupancy still allowed on the track. Vehicles whose requested periods of use are in the same segment do not overlap, do not interfere. Therefore, at the allocation of segments only be noted that the maximum Occupancy number for no time exceeded becomes. In order to meet this condition efficiently, working the procedure uses fixed time slices that set the time in even intervals divide. Every vehicle must be for each Requested segment time slices specify the usage period include. The auctioneer distributes the route segments for each requested Time slice to the bidding vehicles in the order of their Bids (from the highest started). When the maximum occupancy number for a route segment is reached is, can no more vehicles will be allowed on this segment, even if they have requested this.

The The result of an auction round is that each Vehicle receives part of the requested route. In the best case, it receives the entire route, in the least favorable at all no segment of it. In the middle case, the vehicle receives a route with "holes" of segments where it has not been awarded the contract. The following auction rounds have to each vehicle serve to close these "holes." A vehicle may have a Route only from the starting point to the first point of a segment use that did not get it. Behind "holes" lying segments are worthless Everybody has to Vehicle try to get one as possible long coherent piece behind to secure the starting point. This is in particular bidding behavior to take into account.

After this the vehicle coordinator has determined the end of the auction i.e. does not initiate any new auction round, get all the vehicles the contract for the use of the first part of the route, up to the last segment, up to which the vehicle has consistently bought the route runs. Behind the first gap lying auctioned segments are therefore no longer considered, since they are for the vehicle is worthless. The vehicle retains the budget for these segments for future auctions. Ideally, a vehicle gets the entire route, in the worst Case an empty route.

To The acquisition of a route or a route part gives each vehicle a tolerance measure in Percent by which the quality of the route traveled deteriorate may, without that Vehicle at a new auction to buy a better one Would like to take the route. If the current traffic situation and thus the evaluation of this Route deteriorated by more than the tolerance level (due a fault as a traffic jam or total blockage), then the vehicle reports its needs at the vehicle coordinator.

The Process of the invention receives a significant added value through the implementation of individual Auction strategies for the vehicles. These auction strategies contribute to the autonomy of the individual vehicles invoice. In principle, any auction strategies integrable in the AVIR prototypes.

The following should be considered by each tenderer:
In order to make an offer for a track, offers must be submitted for all sections of the track concerned. For this it is necessary that the bidder calculates in which time interval he needs the respective section of the route. This invoice must take into account when the section of the route is used as soon as possible and when it will be left as late as possible. In order to make the system less susceptible to possible displacements at later points in time, tolerances should be expected right from the start, ie the required time interval should not be calculated too sharply.

The types of bidders can be distinguished according to the following criteria:
The first criterion is the individual assessment of the route quality: There are three basic types according to this criterion: one prefers the shortest route geographically, the second the fastest route and the third the route with the lowest minimum bids.

The second criterion is the strategy of how the bids within the be distributed to the favored route: here there is the guy who the budget he total for the route is ready to be proportional to the individual segments distributed to the minimum bids. Another guy shares the route in two different sizes Split parts and put more weight on the first part. The size of the two Parts and the proportion of greater weight are determined by individual parameters.

One Another criterion is the reaction to the outcome of the previous one Auction Round: During the one type in case of failure immediately tried on the next best Dodge route, the second type redistributes the budget by he offers more to the segments he does not get last time Has. Therefor offers a combination with the type from the second criterion that's a bigger weight on a part of the route lays: It increases the weight of the part in which there are more non-received segments. A third guy gives in the beginning only the minimum bids and increases targeted the offers for the segments he did not get. For this purpose, additional segments can be reached shortly across from Priority segments are prioritized. A fourth Type offers initially only with part of the budget on the best route and increases the bid if necessary. This weighs He starts between the alternatives to raise the bid for the best route or an equal size offer for the next best Route to make.

One Another criterion is the degree of tolerance for the deterioration of the currently traveled route: This degree of tolerance is individualized by each vehicle after receiving a route. Every time the tolerance threshold is exceeded, the vehicle reports to the vehicle coordinator the need for a new one Auction. Here is a distinction between types, always choose the same tolerance measure and Types that the tolerance of the quality dependent on the auctioned route make: If the auctioned route is not the one originally favored is, then the maximum tolerated deterioration of this route significantly lower than the favorite route.

Finally there it still the criterion to what extent precaution for future disturbances operated becomes: the one type consumes all the available Budget in each auction while the second guy starts with only part of his budget and only with increasing proximity to the desired Destination a growing share of his budget.

in the Principle can give each vehicle exactly one type according to each of these criteria be assigned.

The method of the invention can also be applied to the collection of road tolls.

This is it that makes sense the vehicles may continue to move freely in traffic. Indeed should then use a route that is not auctioned has been auctioned, cost more than if they are in the auction process would have been bought. Otherwise there would be there is no reason for the vehicles to participate in the auction process.

A already practiced in the US use of the method exists at the Toll Lanes: These are individual lanes, in contrast Charges apply to the lanes running alongside them. Therefore are They also drive less and promise every user a faster Through. A problem for operators and planners of Toll Lanes is that the fee is reasonable relationship to their benefit. The use of the method of the invention solves this problem for market economy Wise.

When submitting road tolls, the respective fee model that the legislator has provided or approved must be taken into account. The method of the invention can be used in different variants of fee models:
The first variant refers to a basic model that is already used in various countries today: This model provides that the route charges depend only on the type of vehicle and at most from the time of use and otherwise are fixed. The method of the invention can be used in this case for the mere coordination of the vehicles in order to achieve a better utilization of the traffic network and to reduce congestion. A prioritization of the vehicles can be done by allocating different budgets. These budgets are to be understood purely fictitious in this variant and give the vehicle only a certain strength to enforce its reservation request, since the actual fee to be paid is already set in this model. The advantage of soft and not absolute prioritization already comes into play in this variant.

In a second variant, the basic model can be extended: It is for each course and each time a charge corridor is set, i. e. one given vehicle must for the use of Route at a given time an amount between one minimum fee and pay a maximum fee. The actual depends on the amount to be paid from an auction in which the vehicle participates. To this Way, the fee depends within the corridor according to the demand and is due to the cap between minimum fee and maximum fee nevertheless calculable.

In a third variant could the routes are awarded without capping the fees. These Variant would be purely market-based, i. Each participant can communicate with them accordingly buy a lot of money each way. The predictable price for a route But it is difficult to calculate in this way. Notional salary payments At the beginning of an auction are then no longer useful.

The Advantages of the method of the invention correspond to those of the market economy opposite the Planned economy in economic scenarios. Of particular note is the superiority the market economy approach, if spontaneous changes occur in the environment (here for example: congestion, new transport request at the user). A market-based approach also increases customer satisfaction expect, because the customer remains autonomous in his decisions and can prioritize in compressed traffic Dimensions (each after tracked variant) buy.

One Acceptance problem may be that the customer generally unwilling to pay more than the official one Base fee. this can be countered by the second variant with the fee corridor: It would be awkward, the fee as a basic fee (starting from the minimum price) with a mark-up surcharge to sell. Psychologically better is the fee as a basic fee (starting from the maximum price) with Surrender discount for sale. In this reading, a participant receives at auctions a discount that depends on how desired the distance at the time of use is: if the track is little asked, the discount is maximum. If the track is in high demand is, there may be no discount. In this way, the willingness to cooperate at a traffic coordination with a financial incentive.

With For example, aid of the method of the invention may also be problem solved be the transport company affected by it at the moment heavily loaded, namely the incalculable waiting time at border crossings, e.g. to Poland. For this one would have to the allocation of check-in slots at border crossings to the allotment of the previous ones Include motorway routes. With the help of an auction-based procedure can on official and for that all traceable way through targeted use of funds the waiting time can be reduced, a condition at the moment at best can be achieved by non-legal means.

Further advantageous embodiments are an extension of all the embodiments described above:
In reality, the transport networks are quite large. Particularly with a fine granularity of path elements, very complex models of traffic networks are created. To be able to master this complexity better, the entire traffic network is divided into a set of so-called regional network nodes, which are connected to each other via ports in an advantageous embodiment of the method.

The Division of time into discrete time intervals is still global Valid throughout the transport network. Otherwise, each regional traffic node will be covered by the procedure treated its own transport network, i. he has his own auctioneer, and the coordination of traffic control takes place only within the same regional transport node. If a transport unit has a Port wants to drive from one traffic junction to the next in the next Traffic node treated as a new transport unit, which as Start way element the common port with the previous traffic node has and as the start time interval the previously planned arrival time interval the for the planned traffic use of the previous traffic junction. If the planned road use due to a disturbance in the previous network node changes, so must Vehicle in the next Log traffic node with the new data.

The entire method with all advantageous embodiments sets it near, the individual data processing units for the transport units, path elements and the auctioneers and regional network nodes as software agents to program. In this way, the data processing units actually distributed locally and communication using sophisticated agent techniques carried out smoothly become.

It but it should be noted that a local Distribution of data processing units not necessary for the functional ren of the method. It is also a virtual distribution of this Units on a single machine possible.

Claims (34)

Method for controlling transport units in a traffic network, - the having interconnected path elements, with one for each Path element predetermined maximum number of transport units be occupied, the predetermined number during the course of the process is variable, - the Time is divided into discrete time intervals and each transport unit begins its journey at any time interval - each Path element with a minimum fee is occupied, to be paid by a transport unit for the time interval is, to which it occupies the path element, - each transport unit Budget of fee units to disposal is asked, and - the Transport units themselves a way of using an auction mechanism bid. Method according to claim 1, characterized that the Transport units are divided into categories for which different Budgets and various strategies for the use of budgets defined be, and For all transport units, for they do not explicitly have their own budget or strategy has been defined, the budget and the strategy of the associated category is used. Method according to one of claims 1 or 2, characterized that A data processing unit acting as an auctioneer, following a specified auction mechanism the use of path elements for different time intervals to different transport units under the proviso forgives that no way element under one for that fixed minimum fee is assigned and each path element in a time interval only with the maximum number of transport units is occupied, - the auction mechanism includes several auction rounds, in each auction round each transport unit participating in the auction participates a public offer in which there is a required amount of path elements at specific time intervals and offers fees, - every transport unit, who won the contract for has received the use of the path elements last for it offered fee is subtracted from his budget, and - each transport unit one Local processing unit is assigned the iterative calculations of the following kind: 1. Determination of a new path w from Output waypoint to destination waypoint, 2. Determination of required Time intervals for the path elements of a given path w, where also waiting times permissible are. 3. Calculation of an offer for all path elements of a path w under proviso needed in step 2 Time intervals, where the offer for each path element, the minimum fee and the exceed previous highest bid must, under the proviso that for the transport unit fixed budget is not exceeded becomes. A method according to claim 3, characterized in that in summary of steps 2) and 3) a Path utilization corresponding to the path w calculated in step 1 by generating a binary search tree whose nodes are pairs of the form (path element of w, set of time intervals) whose root node is the pair (output path element, start time interval) in which two successor nodes (f ', T') and (f '', T '') are generated at node (f, T), where f 'is the path element following w at w and T' is the set of Time intervals are that is needed if after the last time interval from T immediately the path element f 'is driven and f''is equal to f and T''consists of the time interval that follows the last time interval of T, and where the edge , which connects a node with a successor node (f ', T') or (f '', T ''), is marked with the costs incurred by the auctioning of f 'or f''for the time intervals from T 'or T''arise when the minimum fee and the previous maximum bid for f' or f '' at the relevant Zei intervals are exceeded, and after a search algorithm for shortest paths the most cost-effective route from the root to a node (f, T) is found, in which f is the destination path element, and the nodes of the path calculated in this way in step 2 ) to calculate the required time intervals for the respective path elements and the edges indicate the offers to be calculated in step 3. Method according to claim 4, characterized that already while the calculation of the search tree physically impossible solutions are excluded, by the latest possible time interval, to which the transport unit must leave a segment as additional Mark is placed in the search tree node and child nodes, which correspond to a waiting action, are formed only if their time interval is not after the time interval of the extra Marking the predecessor node lies. Method according to one of claims 3 to 5, characterized that  the Auction ends when all transport units become one Have used auctioned away no path element at the same time more transport units is occupied than the given maximum number pretends, and no transport unit offers more for another Road use gives. Method according to one of Claims 1 to 6, characterized that each Path element is assigned to a local data processing unit, who monitors, that yourself the predetermined maximum number of transport units or others Do not change the parameter and in case of change Notify all vehicles that want to use this path element and of the change of the parameter could be affected Message, and each transport unit is a local data processing unit is assigned, which monitors whether the intended use of the route can be realized, and in case that she can not be realized, the beginning of a new auction triggers, or in a recent auction, a new way use auctioned, with new budgets at the beginning of a new auction distribute the vehicles in the network at the beginning of the auction become. Method according to one of claims 3 to 7, characterized that the Strategy determines when a new route should be calculated and when a new way use for an already calculated Way and an associated Offer should be calculated. Method according to one of claims 3 to 7, characterized that the Strategy determines when a new road use should be calculated and when to a previously calculated Wegbenutzung a new offer to be calculated. Method according to one of claims 8 or 9, characterized that for each category of transport units factors n and m (n <m) are defined and a Transport unit after calculation of a new route and a designated path Wegbenutzung as long as exclusively tries this way use by outbidding all competitors until the fee to be paid is n times the fee charged for the first time, and then resort to another way of using the same route, the for a later Time interval begins, and a transport unit only then consider a new path when the fee to be paid for the other Distance of use is m times the fee calculated for the first time. Method according to one of claims 1 or 2, characterized that the fee for use the way element not for each time element is collected separately, but flat rate for all time elements of use is charged. Method according to claim 11, characterized in that that the maximum occupancy number of a route element with transportation units depending on the time is that a transport unit needs at least to the path element to cross, and that one Function is used in the context of each path element of this context is shown explicitly. Method according to one of claims 11 or 12, characterized that a Data processing unit R is used, which is able to paths to calculate from a given start to a given destination, each Path element is assigned a local data processing unit Q, the current track quality of the path element in dependence from the present present load with transport units, and - in case the method of claim 12 - the functional relationship between maximum occupancy number and minimum traversing time to disposal to Each transport unit is a local data processing unit T is assigned the budget for the transport unit and manages a specific auction strategy, a data processing unit K is used, the various transport units located in the transport network coordinated a data processing unit A acts as an auctioneer, the after release an auction according to one fixed auction mechanism to use the path elements different transport units under the proviso forgives that no path element under that fixed minimum fee is awarded and each way element only with the maximum number occupied by transport units, where each unit T in dependence of its auction strategy and depending on the current route quality of the route elements a path that the transport unit is currently using or that it is considering pulls to use, announces the need for a new auction at K and submit offers to individual way items that at least that way are high as the respective minimum fee, and K at A an auction after a K predetermined heuristic triggers that may depend on how many units T have declared the need for an auction, and the transport units the offer prices with which these have bought a way from the end of the auction, from the entire to disposal standing budget. Method according to claim 13, characterized in that that the Transport units due to the current traffic situation and - optionally - with the help forecasting tools the expected period of use for each Compute path elements lying before them, this by a safety period which are individually determined by each transport unit may, and submit the calculated total period in the offer, in which the auctioneer A assigns the way elements under the proviso that to none When a path element is occupied by more transport elements as specified in the maximum number. Method according to one of claims 13 or 14, characterized that the Transport network may also be occupied by transport units the process has no influence and on each path element only so many transport units through the process be distributed that the sum the transport units distributed through the process and the current ones Number of non-influenceable Transport units or - optionally - the forecast Number of non-influenceable transport units in no case exceeds the maximum occupancy. Method according to one of claims 13 to 15, characterized that the Data processing unit K after the submission of offers by willing to transport units and before the release of the Auction still determined other auction participants, not from have subscribed to the need for a new auction, which is but to be on a path for which an offer has been made, whereby - in case of the method according to claim 14 - the requested period of use with the predicted usage period of en route Overlap transport unit got to. Method according to one of claims 13 to 16, characterized that a Auction in one or more auction rounds expires and each auction participant before submitting an offer for one decides the path calculated by the data processing unit R. and then for all the elements of this path make an offer and in everyone Auction round the offers of all auction participants in parallel become and the data processing unit K after each auction round Determines which participants want a new auction round and decides the continuation or abort of the auction after a heuristic, the depends on how many auction participants are allowed to make a new auction round have registered. Method according to one of Claims 13 to 17, characterized in that the auctioneer distributes the route elements to the auction participants in each auction or round of auctions, by submitting for each route element the transport units which have placed an offer on this route element in the order of their bid - from starting highest - getting the way element allocated until the maximum occupancy number is reached, and each transportation unit receiving a way element according to this method receives it at the price equal to the highest offer among the providers who have not received this way element anymore , or if all providers the way element he keep at the price of the minimum bid. Method according to one of claims 13 to 18, characterized in that each transport unit T, which has auctioned off parts of a path at the end of an auction, only the first part, starting from the starting point of the path to the last in succession path element W _l , the T has auctioned, gets allocated for use and all behind the first path element W _{l + 1} , which has not bought more T, lying path elements W _i does not get allocated for use, regardless of whether T has purchased these path elements W _i or not , Method according to one of claims 13 to 19, characterized that each Transport unit at the beginning of an auction in addition to its budget all fees gets refunded for it Had to pay path elements, which it did not use at the beginning of the auction. Method according to one of claims 13 to 20, characterized that the Auction strategy of the transport unit for each route between takeoff and destination, which was named by the data processing unit R, the path quality individually rated and then selects the best path, taking the properties geographical length, current usage duration and sum of minimum fees for the Assessment can be considered. Method according to one of claims 13 to 21, characterized that the Auction strategy of the transport unit a way between start and Select destination as best, For this Path makes a bid and the bids on the individual way elements differently weighted distributed on a front and a back part, the relative number of path elements of the front part and the proportion of bids for the front part of the bid for the entire path is given by individual parameters. Method according to one of claims 13 to 22, characterized that the Auction strategy of the transport unit from the exit of the last auction round depends by the transport unit for the case that she not all path elements of the desired Got way, either evades another way or the given bid is redistributed by path elements that form the transport unit in the last auction round with fewer funds could have bid on Path elements they do not get in the last auction round Has, or in the first auction round only the minimum bids and allocate more funds in the following auction rounds the way elements they used in the last auction round did not get or in the first auction round only one Part of the total available standing budgets for the wished Way and in the following auction rounds either the bid for the increased all the way or the same size, but it does for one other way. Method according to one of claims 13 to 23, characterized that the Auction strategy of the transport unit an individual tolerance measure for a Way sets the transport unit in the last auction has bought where the tolerance measure determines from which percentage Deterioration of the current route quality of the auctioned route Need a new auction is logged. Method according to one of claims 13 to 24, characterized that the Auction strategy of the transport unit a certain share of the disposal standing budgets in principle withhold to future Auctions a bigger budget as the new salary available to have, with the withheld May depend on the relative distance to the destination. Method according to one of Claims 11 to 25, characterized that it as a billing procedure for charging road traffic is used, the road vehicles in the process transport units, of which a fee for the use of the road will be charged. Method according to claim 26, characterized that the Vehicles are free in the traffic allowed to move, but for the use of path elements that they did not purchase, a higher one fee have to pay, as any vehicle that has bought it. Method according to one of claims 26 or 27, characterized that the Path elements are individual tracks of roads that require a fee are. Method according to one of Claims 26 to 28, characterized that it for a fee model is used in which the real fee to be paid for a way element only of the vehicle type and the usage time may depend but not by the number of other vehicles, the same way element want to use at the same time the method used Budget is fictitious and only for prioritizing the vehicles serves. Method according to one of claims 26 to 28, characterized in that it is used for a charging model in which the real fee for a path element depends not only on the type of vehicle and the time of use, but also on the number of other vehicles wanting to use the same path element at the same time, with a fee corridor for each path element between a minimum charge and a maximum charge and no vehicle - irrespective of its offer - has to pay more for the use of a way element than the maximum charge. Method according to claim 30, characterized that the base fee for a Path element of the maximum fee and each vehicle the difference to the auctioned price is granted as a discount. Method according to one of Claims 26 to 28, characterized that this Budget used in the procedure corresponds to the real budget available and no salary is paid at the beginning of an auction. Method according to one of claims 1 to 32, characterized that the entire transport network in order to reduce the complexity in regional Network node is subdivided, which in turn is constructed like traffic networks are and the over Ports are connected to each other, and every regional network node has his own auctioneer and the control of the transport units made independently in the different regional network nodes and the transport units that are different from a regional network node to move to another, after the purchase of a Wegbenutzung in the first network node with the path element as output element and the time interval Log in as the start time interval in the second network node, which is them the auctioned off for use the transitional port assigns between the two network nodes, and every transport unit at a change of the path usage assigned to it by the first regional hub, the the transitional port concerns, new at the second regional network node for assignment a Wegbenutzung logs. Method according to one of claims 1 to 33, characterized that the Data processing units for the transport units and track elements as well as for the auctioneers or regional network nodes are programmed as software agents.