Method for coordination of competing processes or for control of the transport of mobile units within a network

• US 8,103,434 B2
• Filed: 05/10/2006
• Issued: 01/24/2012
• Est. Priority Date: 05/17/2005
• Status: Active Grant

• Current Assignee: Eth Zuerich
• Original Assignee: Eth Zuerich
• Inventors: Helbing, Dirk, Lämmer, Stefan
• Primary Examiner(s): Trammell, James
• Assistant Examiner(s): Marc, McDieunel
• Application Number:
  US11/914,759
• Publication Number:
  US 20080235398A1
• Time in Patent Office: 2,085 Days
• Field of Search: 701/117, 701/118, 701/211, 701/209, 701/202, 701/210, 701/182, 701/212, 340/910, 340/911, 340/917, 455/575.9
• US Class Current: 701/117
• CPC Class Codes: G05B 13/024 in which a parameter or coefficient is automatically adjusted to optimise the performance
• CPC Class Codes: G08G 1/01 Detecting movement of traffic to be counted or controlled G08G1/07 - G08G1/14 take precedence
• CPC Class Codes: G08G 1/07 Controlling traffic signals

Abstract

A method for coordination of concurrent processes or for control of the transport of mobile units within a network, wherein a) the control of the network occurs in a decentralized and self-organizing manner in the controllers of node points or local defined sub-networks, whereby the control units of adjacent node points or sub-networks are connected to each other for data exchange of, b1) data from prediction models for local process sequences at each node and/or data from prediction models for the local process sequences of adjacent nodes and/or b2) data from data recording elements of each node or the boundaries associated therewith and/or data from data recording elements of adjacent nodes or the boundaries associated therewith, c) local simulation and optimization of switching the controller to establish the performance of the nodes or sub-networks with regard to the buffer capacity of the boundaries based on models for short-term predictions with fixed switch states for adjacent nodes, are applied.

Claims

1. A method for coordination of competing processes or for control of a transport of mobile units within a network which has nodes and edges, with the network being subdivided into core areas n, each of which has an associated edge area so that the core area and the associated edge area each define a cohesive subnetwork of nodes and edges, with the edges having a limited buffer capacity for units to be received and with the nodes or edges being equipped with data acquisition elements, and with the nodes being equipped with a limited dispatch capability for units to be controlled and being equipped with control units for which a (switching) state of dispatch of mobile units, a (switching) state of non-dispatch and, between them, a changeover switching state are provided, wherein:

a the network is controlled on a decentralized and self-organizing basis in control units of the node points or locally limited subnetworks, with the control units of adjacent node points or subnetworks being connected to one another for data interchange,

b1 data from prediction models of local process procedures at the respective node, data from prediction models of local process procedures at adjacent nodes, or data from both prediction models of local process procedures at the respective node, and data from prediction models of local process procedures at adjacent nodes, or

b2 data from data acquisition elements of the respective node or of the edges connected to it, data from data acquisition elements of adjacent nodes or of the edges connected to them, data from both data acquisition elements of the respective node or of the edges connected to it and data from data acquisition elements of adjacent nodes or of the edges connected to them, or a combination of data according to b1 and b2

c are used for local simulation and optimization of switching sequences of the control unit in order to determine the performance of the nodes or subnetworks taking into account the buffer capacity of the edges on the basis of models for short-term predictions with fixed assumed switching states of adjacent nodes, with

c1 a plurality of high-performance switching strategies being produced for the subnetworks by a combination of high-priority switching sequences for the relevant individual nodes, and

c2 a test of the high-performance controllers being carried out in the subnetworks, and, subsequently,

c3 that control strategy which has the best performance being selected in the respective core area of the subnetwork and being converted to corresponding switching states for the relevant nodes.

2. The method as claimed in claim 1, wherein destination and optimization criteria for adjacent nodes included in the local optimization in method step c are weighted as a function of the position in the respective local optimization region (subnetwork).

3. The method as claimed in claim 2, wherein the weighting is carried out in the decentralized control method by maximizing the expression

4. The method as claimed in claim 1, wherein, for the simulation process in method step c:

d1 when there is little demand for the mobile units at the nodes, a single dispatch, with equal priority, of the units is carried out without any waiting times, with the number of dispatched units being proportional to the average number of arriving units,

d2 when there is a high demand with unavoidable waiting times, the units are dispatched in groups by means of relatively long switching phases based on the principle of minimizing the switching losses, with

d3 the process of minimizing the switching losses being combined as required with the prioritized dispatch of units close to the destination in order to quickly remove them from the network,

d4 switching time losses being more than compensated for by premature switching in favor of synchronization of adjacent dispatch switching states by means of a predicted higher dispatch capability at adjacent nodes,

d5 buffer capacity being reserved as required in the edges following the node, with the buffer capacity provided per edge being adaptively matched,

d6 prioritization, on the basis of which the units are dispatched at the nodes, being carried out on the basis of the characteristics of the units, and in that

d7 the maximum duration without any dispatches being restricted by setting of quota.

5. The method as claimed in claim 1, wherein the local (decentralized) optimization uses a database-assisted method in which a number of high-performance control strategies are stored as a function of parameters which characterize the situation in the respective subnetwork and are raised as part of the method.

6. The method as claimed in claim 1, wherein the quota are set on the basis of criteria of user or system optimality.

7. The method as claimed in claim 1, wherein the required model parameters are determined by a combination of a plurality of alternative simulation and measurement data items.

8. The method as claimed in claim 1, wherein, in the process of estimating the predicted dispatch times of a unit, measured values, short-term prediction, or both measured values and short-term prediction are more strongly weighted on close-by edges, while in contrast historical measured values are more strongly weighted for comparable situations on remote edges.

9. The method as claimed in claim 1, wherein:

if the measurement data according to method step b2 is missing or implausible, the missing or implausible data is replaced by a simulation or estimation model, or

if the simulation or measurement data according to method step b1 is missing or implausible or if data according to both method step b2 and method step b1 are missing or implausible, the process is controlled on the basis of fixed predetermined control.

10. The method as claimed in claim 1, wherein units which are associated with incompatible dispatch processes, that is to say dispatch processes that are subject to conflict, are dispatched individually at appropriate times in parallel with the dispatch of units in groups.

11. The method as claimed in claim 1, wherein data relating to unit or process destinations is interchanged with the control units for the nodes for prioritized dispatch of units or processes that are close to the destination.

12. The method as claimed in claim 1, wherein continuous signals or control actions are used to coordinate concurrent processes or to control the transport of mobile units.

13. The method as claimed in claim 1, wherein the method is used to optimize a network structure by variations of nodes and edges.

14. The method as claimed in claim 1, wherein this method is combined with a method for optimizing the turning probabilities.

15. The method as claimed in claim 1, wherein the method is combined with a method for optimizing the dynamics of the individual units along the edges, with the units receiving information, instructions or both information and instructions from the signal controller.

16. Use of a method as claimed in claim 1 for control of a traffic network, with the nodes in the network representing the crossings and the control units representing the traffic lights, and with the edges representing the road sections and the mobile units representing vehicles or others involved in the traffic network.

17. Use of a method as claimed in claim 1 for controlling a production process, with the nodes representing the production machines or productive units, the edges representing the transport paths or intermediate stores between them, and the mobile units representing the products.

18. Use of a method as claimed in claim 1 for controlling logistics of goods transport, with the nodes representing the transfer points, the edges representing the transport paths, and the mobile units representing the transport goods.

19. Use of a method as claimed in claim 1 for coordination of organizational procedures, with the units corresponding to the processes and the nodes corresponding to workers, while the edges describe the management or feed path of the processes, and with the controller corresponding to the prioritization decisions made.

20. Use of a method as claimed in claim 1 for handling program procedures, with the nodes corresponding to the program modules which process data, with the data being passed on between the program modules defining the edges of the network of successive processing processes, and with the controller being used for coordination or harmonization of the processes.

21. Use of a method as claimed in claim 1 in the field of biotechnology for controlling the distribution process of substances or active substances in cells, tissues or organisms.

22. A traffic system having a multiplicity of crossings, road sections, data acquisition elements and traffic light installations with control devices which are suitable for carrying out a method as claimed in claim 1, wherein the crossings have associated traffic light installations and sensors, with the control devices being connected to the sensors in order to receive data, and in that the control devices for adjacent traffic light installations are connected to one another in order to interchange data.

23. The traffic system as claimed in claim 22, wherein the data link between the sensors and the control devices for the traffic light installations as well as between the control devices for adjacent traffic light installations is wire-free.

24. The traffic system as claimed in claim 23, wherein the data link is in the form of a wireless LAN (WLAN), as Bluetooth, as an infrared link, as a radar signal, as a laser link or as a variant thereof.