WO2001050666A2 - Programmateur de ressources exerçant un effet positif sur les couts et procede correspondant - Google Patents

Programmateur de ressources exerçant un effet positif sur les couts et procede correspondant Download PDF

Info

Publication number
WO2001050666A2
WO2001050666A2 PCT/US2000/035090 US0035090W WO0150666A2 WO 2001050666 A2 WO2001050666 A2 WO 2001050666A2 US 0035090 W US0035090 W US 0035090W WO 0150666 A2 WO0150666 A2 WO 0150666A2
Authority
WO
WIPO (PCT)
Prior art keywords
train
cost
resource exception
resource
schedule
Prior art date
Application number
PCT/US2000/035090
Other languages
English (en)
Other versions
WO2001050666A9 (fr
WO2001050666A3 (fr
Inventor
Michael S. Crone
Original Assignee
Crone Michael S
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from US09/476,615 external-priority patent/US7092894B1/en
Application filed by Crone Michael S filed Critical Crone Michael S
Priority to AU24529/01A priority Critical patent/AU2452901A/en
Priority to CA002397123A priority patent/CA2397123A1/fr
Publication of WO2001050666A2 publication Critical patent/WO2001050666A2/fr
Publication of WO2001050666A3 publication Critical patent/WO2001050666A3/fr
Publication of WO2001050666A9 publication Critical patent/WO2001050666A9/fr

Links

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B61—RAILWAYS
    • B61L—GUIDING RAILWAY TRAFFIC; ENSURING THE SAFETY OF RAILWAY TRAFFIC
    • B61L27/00—Central railway traffic control systems; Trackside control; Communication systems specially adapted therefor
    • B61L27/10—Operations, e.g. scheduling or time tables
    • B61L27/12—Preparing schedules
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B61—RAILWAYS
    • B61L—GUIDING RAILWAY TRAFFIC; ENSURING THE SAFETY OF RAILWAY TRAFFIC
    • B61L2205/00—Communication or navigation systems for railway traffic
    • B61L2205/04—Satellite based navigation systems, e.g. global positioning system [GPS]

Definitions

  • the present invention relates to the scheduling of movement of plural units
  • Trains may also be hauling empty cars back to a terminal for reloading and may be carrying equipment or
  • Freight trains may also be
  • the railway system must both attempt to schedule such trains in a way that the
  • the diverted train may then be
  • a main track 10 may include a main track 10, a side track 20 which is selectively utilized through switches
  • the switches may be manually operated or may be remotely operated through a
  • the HUT 24 may receive signals from track sensors 26 which indicate the presence of a train on a
  • the train system may also include aspects 28 which are illuminated lamp
  • train detection sensors 26 operate along a length of track which may be
  • sections of track are isolated into separate segments by discontinuing the track for a brief
  • each segment of track is electrically isolated
  • a voltage differential is applied between the two rails of a track and when a train
  • the metal wheels and axle Of the trains serve as a conductor electrically
  • the HUT 24 is able
  • the HUT 24 may send information regarding various of the conditions supplied to
  • the present systems as described above, provides positive separation between trains so
  • meeting trains is known. For example, if it is known that a train travelling thirty miles an
  • ATCS has been designed and includes transponders, locomotive interrogators,
  • transponders are placed between or near
  • the odometer error provides an uncertainty as to the
  • transponders along the entire railway system may substantially increase maintenance costs
  • transponders are relatively sensitive electronic elements in a harsh environment.
  • a stringline plots time along one axis and track miles or
  • the grid of Figure 2 for example, runs from 5:00 a.m. on
  • the stringline for a train appears as
  • train 11 was sided at Brovo for nearly two
  • train 88 for example, spent almost two hours of a five hour
  • train G7 With more precise knowledge regarding the location of the trains, train G7 may have
  • the train may be sided for a period of time, i.e., the meet and passing was not put into
  • siding should be used to allow the trains to pass, and to set the appropriate switches and
  • the dispatcher generally controls only a portion of the rail system and
  • a schedule should involve all elements or resources that are necessary to
  • the main lines can be used at capacity.
  • a first step in providing a precision control system is the use of an optimizing
  • boundary conditions can include
  • extrinsic traffic which in the U.S. is most often passenger traffic
  • the present invention may be structured to facilitate the use of new boundary
  • the present invention determines a movement plan
  • present invention incorporates into the schedule the very fine grain structure necessary to
  • Such fine grain structure may include
  • the present invention provides the movement plan to the persons or
  • the movement plan can be provided merely to the dispatching
  • the movement plan may be provided to the locomotives so that it can be implemented by the engineer or automatically by switchable actuation on the
  • the disclosed system and method may be viewed as a transportation system in
  • the capacity of the given rail system is the minimum spacing of the trains and the relative
  • the train movement planning system disclosed herein is hierarchial in nature in
  • This hierarchial process means that the solution space over which
  • the network at a predetermined time may be set as one of the boundary conditions on the
  • track at a particular time may be set as a boundary condition and trains may be moved
  • base may be that so many hours of maintenance activity on a given section of track must
  • the scheduler may be allowed to schedule that activity in concert
  • the present invention is that it has continuous monitoring of anomalies as they occur
  • handling logic element which determines at what level the anomaly may be resolved.
  • anomaly resolution or exception handling process can be involved in various levels of a
  • priority must be tempered and the priority function must be delayed
  • Incremental cost can be fuel cost, hourly cost of personnel, hourly use cost of locomotives
  • the plan must include nonlinearities in the incremental costs to allow for the fact
  • step function or as a slope.
  • a true optimization plan is one whereby the variables including the assignment of
  • the four late train may be given a much lower
  • total cost includes the operating costs such as fuel and rolling stock utilization as well as
  • the present invention provides a novel method and apparatus for
  • Figure 1 is a schematic block diagram of the prior art systems.
  • Figure 2 is a pictorial depiction of a prior art stringline used in the scheduling of an
  • FIG. 3 is a functional block diagram of the system of the present invention.
  • Figure 4 is a functional block diagram of the system wide planner or order
  • FIG. 1 is system flow diagram of the implementation of the resource scheduler of
  • Figure 6 is a functional block diagram of the movement planner portion of the
  • Figure 7 is a functional block diagram of the physical model of Figure 6.
  • Figure 8 is a schematic illustration of system operation.
  • Figure 9 is a pictorial illustration of the multilevel abstraction of the three
  • Figure 10 is a functional block diagram of the train controller of Figure 3 as may
  • Figure 11 is a functional block diagram of a portion of the train controller of
  • Figure 12 is a graphical representation of the ideal trajectory of the target resource
  • Figure 13 is a graphical representation of the simplified trajectory of the target
  • a system wide planner or order scheduler 200 may include a system wide planner or order scheduler 200, a
  • planner/dispatcher 204 a safety insurer 206 and a train controller 208.
  • system wide planner 200 is
  • planner 200 develops a coarse schedule for the use of the various resources and passes
  • the planner/dispatcher 204 receives the
  • the movement plan determines a detailed schedule of the resources termed a movement plan.
  • plan may then be used by the dispatching portion of the planner/dispatcher 204 to be
  • the movement plan developed by the planner/dispatcher 204 may ' be checked by a
  • safety insurer 206 to verify that the movements being commanded by the
  • the planner/dispatcher 204 may also
  • the signals to the track elements 210 may be verified for safety by the safety
  • parameters include, for example, the grade of the track, its curvature and slope, and the condition of the track bed and rails.
  • the system wide planner 200 is able to generate a coarse
  • all levels of the system may include train handling constraints within
  • Train handling constraints include experiential and other factors by
  • the planner/dispatcher 204 of Figure 3 has two processes: a planner/dispatching
  • the planner/dispatching function is responsible for the
  • the movement plan is a time history of the position of the trains throughout the
  • the movement planner takes into account
  • the movement planner takes into account the speed changes and/or time
  • the movement planner accounts for the fact that the train may have to
  • the movement planner generates the exact trajectory which the train is expected to follow.
  • the models of the prior art may model the travel between two
  • Fixed block rules reflect the segmentation of tracks into fixed blocks or
  • the block size was set at the distance that the
  • the separation between trains can be made smaller than in the fixed
  • the system of the present invention is not based on a "worst case” braking
  • selective portions of the movement plan can be
  • the movement plan can be configured to the various track elements (switches, signals, etc) as called for in the movement plan.
  • the movement plan can be
  • Both the movement plan signals and the track force controlling signals may be identical
  • the safety insurer may be any suitably
  • the train controller 208 would not be able to implement the plan and could
  • planner or order scheduler 200 may include an extent of planning determiner 304, an activity identifier
  • the order scheduler 200 may also include a constraint based inference engine comprising
  • a new order for rail service may be applied via an input
  • the order may be any request for
  • rail service may include an origination point, an earliest pickup time at the origination
  • An order may take the form of a request to move a specifically loaded train from
  • the sequence of track resources is, of course, dependent upon the selection of a route if alternative routes are available.
  • the extent of planning determiner 304 also receive on an input terminal 306 the
  • a resource may be any entity which may be scheduled
  • the extent of planning determiner 304 may also receive any schedule exceptions
  • a schedule exception may be any previously scheduled event
  • the extent of planning determiner 304 may also receive any extrinsic traffic which
  • Extrinsic traffic is any traffic which is not subject to
  • extrinsic schedule exemption for the typical railway freight system may be the inviolate
  • the extent of planning determiner 304 may be any suitable conventional apparatus,
  • the extent of planning determiner 304 provides orders to an activity identifier and sequencer 310 via terminal 312 and the activity identifier and sequencer 310 provides an
  • An activity is an event which requires one or more resources to be assigned for a
  • an activity may be the loading of a train with a bulk
  • the activity identifier and sequencer 310 in turn provides a list of the available
  • the activity identifier and sequencer 310 may be
  • the list of candidate resources ' from the activity identifier and sequencer 310 may
  • the train action effects calculator 318 also provides an input
  • calculator 318 may be any suitable conventional appropriately programmed general
  • composition of the train the effects which the terrain over which the train travels has thereon. While not limited thereto, the effects of terrain on the acceleration and
  • the calculator is provided with the
  • the time interval converter 320 may likewise be suitable conventional general
  • the output signal from the time interval converter 320 may be applied by way of a
  • the interval grouper 324 also receives via
  • the interval grouper 324 may be any suitable conventional general or special
  • the resource scheduler 330 which receives the interval groups also receives by
  • the scheduler 330 receives a signal from the extent of planning
  • determiner 304 indicative of the resources available for the scheduling process.
  • the output signal from the scheduler 330 is the schedule which is also fed back to the extent of planning
  • determiner 304 as discussed below.
  • the resource scheduler 330 may be any suitable conventional general purpose or
  • the resource scheduler 330 is desirably one
  • the extent of planning determiner 304 determines the extent of
  • the extent of planning determiner 304 uses a set of rules defined by standard operating
  • the orders from the extent of planning determiner 304 are received by the activity identifier and sequencer 310 and are used to generate an activity list. For each order, a
  • the activity list includes the
  • sequence of track segments i.e., route which must be traversed in filling the order.
  • Route selection may be based upon cost analysis, upon previously determined company
  • the activity list is, of course, ordered
  • the activity list is supplied to the candidate resource determiner 314. For each of
  • locomotive power may not be able to move over the grade associated with the selected
  • the candidate resource determiner 314 serves
  • this computation may be
  • Loading and unloading tasks may be computed by dividing the capacity of a train
  • This rate may be
  • the time computed for each of the activities on the activity list is adjusted for
  • the time interval converter 320 translates the sequence of activities on the activity
  • action effects calculator 318 for each of the activities identified by the candidate resource
  • interval converter 320 passes a list of time intervals grouped by resource as well as by
  • the interval grouper 324 receives the list of grouped intervals from the time
  • interval converter 320 receives the orders from the extent
  • planning determiner 304 groups the time intervals necessary to fulfill the orders in
  • the interval grouper 324 provides the time intervals
  • Gaps represent the time periods which may be allowed to pass between the
  • gap may be the existence of a siding or other capacity for holding a train for an interval of
  • a gap e.g., one associated with the passage of a train over a section of track to a
  • interval groups are passed to the resource scheduler 330 which also received
  • the resource scheduler 330 thus conducts a search for a schedule which satisfies the resource
  • the search for an acceptable schedule may employ various
  • the interval groups are returned to the interval grouper 324 for
  • the resource scheduler 330 can provide a schedule which meets
  • the display 334 conveniently displays the resulting schedule for user examination.
  • a popular display is a standard string-line diagram used by the railroads such as
  • the resource scheduler 330 performs globally optimized scheduling of train
  • the resource scheduler 330 is implemented in the Harris Corporation
  • This shell provides a Constraint Propagation Expert System (COPES) Shell. This shell provides a
  • the resource scheduler 330 shown in Figure 5 is a UNIX process which schedules
  • Multiple orders may be scheduled either in batch or
  • an order also has a time interval during which the
  • the activity lists can be converted to a sequence of time intervals by incorporating the
  • Focused simulated annealing is a distributed version of simulated annealing written in COPES. It follows the traditional flavor of simulated annealing in the random
  • Variables include starting temperature and the number of temperature reductions
  • Each routine can be used to each trip to make decisions themselves about how useful modifying the current trip (e.g., start time, equipment assigned) would be to the overall situation.
  • Each routine can be used to make decisions themselves about how useful modifying the current trip (e.g., start time, equipment assigned) would be to the overall situation.
  • variable reflecting the importance of the next move of the trip e.g. a larger time range
  • the resource scheduler 330 employs a dynamic, distributed, robust, and efficient
  • the temporal logic also considers constraints such as moving block distances.
  • trip instance such as "tripO_state” is actually composed of trip state variables
  • trip_resource class objects defining the sequence of resources necessary to complete the
  • the time interval converter requests a schedule from the resource scheduler 330.
  • the server o constraint fires and moves this request into the interface state class which
  • This message contains information about the order as described earlier, search goals, and
  • the op_capacity_request constraint generates order class objects for each
  • Control_search initializes the search and annealing parameters and sets up for the
  • the schedule for firing is a discrete-event queue reflected by. scheduled
  • control_search starts another annealing pass with
  • a directed search process is performed to further refine the schedule
  • the move operators is variable depending upon the phase of the search, goals of the search, and their likelihood of improving the solution. At lower temperatures the move operators is variable depending upon the phase of the search, goals of the search, and their likelihood of improving the solution. At lower temperatures the
  • the change_equipment is only fired if it is determined at a low temperature that the train
  • the move_group operator is
  • the scheduling system utilizes a cost reactive resource scheduler to minimize resource exception while at the same time minimizing the global costs associated with the solution.
  • resource exception is the amount of time that two or more resources are in conflict, e.g., the duration of time that two trains are scheduled to be using the same track at the same time.
  • a cost reactive scheduler may be used to develop a schedule by evaluating the resource exception and the cost associated with moves resulting in a schedule which is not only resolvable, but represents a more minimal cost solution. Since the movement planner is designed as a hierarchical system, and the abstraction used for resources and movement by the scheduler leave room for the movement planner to resolve minor conflicts, the resource scheduler does not have to remove all resource exceptions for a successful solution to be found.
  • the cost reactive scheduler comes as close to this solution, i.e., target total resource exception time of 1% of the total unopposed trip time, as possible, without going under it, to minimize the resultant global cost.
  • the cost reactive scheduler is able to achieve this minimum global cost by evaluating the "goodness" of each move in terms of resource exception and cost associated with each move.
  • the cost reactive scheduler does not resolve each scheduling problem in the same way.
  • the cost reactive scheduler initially classifies the set of orders for train resources and then generates a schedule using scaling parameters and acceptance criteria which are dependent upon the classification of the scheduling problem. For example, the cost reactive scheduler may initially classify a set of order for train resources into one of four categories:
  • the scheduler will classify the problem as "Cost Constrained” and will emphasize cost. If it appears that there is excessive slack in the solution space, the scheduler will classify the problem as "Cost Constrained and Resource Unconstrained” and will emphasize reducing cost even more. If the scheduling problem appears to have insufficient slack to achieve the 1% target resource exception, the scheduler classifies the problem as "Resource Constrained” and will emphasize reducing resource exception at the expense of cost. All other scheduling problems may be considered ordinary and have a straight forward solution and may be classified as "Normal".
  • the cost reactive scheduler will determine a scaling parameter which may be applied to the resource exception or the cost associated with each move to emphasize either the resource exception or the cost as a function of the classification of the scheduling problem.
  • the cost reactive scheduler searches for moves that approach the target resource exception of 1%.
  • the cost reactive scheduler accepts a lesser solution, in order to preserve moves for later in the search phase that would not otherwise be available if the 0% resource exception solution was initially accepted.
  • the scheduler may accept a move where the result of the move results in an scheduler (solid line) can be seen varying above and below the trajectory during the search.
  • the effect of the cost reactive modifications is to pull the actual values toward the trajectory.
  • FIG. 13 a simplified piece- wise linear approximation of the idealized trajectory was found to accomplish the desired goal in a more efficient manner than by deriving a polynomial to represent the experimentally determined trajectory.
  • the approximation is shown as a dashed line in Figure 13.
  • the target trajectory is represented by cost emphasis linear phase with zero slope (and an exception value of -30000 seconds in this example), followed by one with a slope like the idealized curve.
  • the maximum number of temperature steps representing the cost emphasis phase is a dependent upon the classification of the scheduling problem.
  • the goal of the cost reactive scheduler is to generate a schedule within this target range - thus insuring a low cost solution.
  • the cost reactive scheduler may receive a set of orders for train resources (the movement of trains or utilization of other resources) which define a scheduling problem.
  • Each order will have a cost function, typically a polynomial equation associated with it.
  • the order may also have a scheduling window associated with it i.e., earliest departure time and latest arrival time.
  • the cost reactive scheduler will classify the scheduling problem as a function of the slack associated with the orders.
  • Slack is the accumulation for all of the orders of the differences for each trip between maximum trip time based on the scheduling window and the minimum trip time based on maximum throttle.
  • the scheduler may compare the slack associated with the scheduling problem with the total trip time for the scheduling problem, and classify the scheduling problem as Cost Constrained, Cost Constrained and Resource Unconstrained, Resource Constrained or Normal as previously discussed.
  • the scheduler should be able to achieve the target 1% resource exception and the problem may be classified as "Cost
  • Another predetermined parameter may be used to determine if a problem is Resource Constrained. For example, if the resource exception time divided by the total trip time is greater than this predetermined parameter, then the resource scheduler may classify the problem as "Resource Constrained".
  • the scheduler may classify the problem as "Normal".
  • the scheduler may perform several other functions before initiating the search phase. As explained above, the scheduler will emphasize costs during the beginning of the search phase ("cost emphasis phase"). The duration of the cost emphasis phase may be based on reducing the resource exception to a specified level, or on a maximum number of moves or temperature steps. For example, the scheduler may determine the initial resource exception value and the initial cost associated with the scheduling problem. The scheduler may then determine the level to which the resource exception value must be reduced in order to stop emphasizing costs. The scheduler may also determine a maximum number of temperature steps for emphasizing costs based on the classification of the problem, with the Cost Constrained problem requiring the most temperature steps and Resource Constrained problem requiring the least temperature steps.
  • the scheduler may estimate target resource exception as a specified percentage of the total trip time by adding a percent to the total minimum trip time and
  • the target resource exception may be updated periodically, e.g., after every 800 temperature steps.
  • the initial scaling parameter may be defined by the COPES database.
  • the search phase may comprise a first phase and a second phase.
  • the normalizing component of the scaling parameter may be determined and updated after every move.
  • the normalizing component may be defined as the ratio of the change in resource exception versus the change in cost.
  • the scheduler may also define the normalizing component as a function of the classification of the scheduling problem. For example, if the scheduling problem is Resource Constrained, then the scheduler may retain the largest change ratio as the scaling parameter. For all other classifications, the scheduler may retain the ratio of the average change in resource exception versus the average change in cost as the normalizing component.
  • the duration of the first phase of the search may be defined by the COPES database , e.g., the initial 100 temperature steps. After the first phase, the normalizing component is no longer updated and remains constant throughout the second phase of the search.
  • the biasing component may be used to force the resource exception toward the target trajectory.
  • the first phase begins the search for a solution to the scheduling problem by making a random move.
  • the resulting resource exception for the problem and the cost associated with the move may be determined by applying the initial scaling factor (for the first move) to the resource exception value and the cost as a function of the classification
  • the cost will be weighted more heavily than the resource exception.
  • the normalizing component from the previous move is used to determine the scaling parameter for the subsequent move.
  • the biasing component may continue to be determined during the subsequent move as a function of the resource exception as compared to the target resource exception.
  • the normalizing component may be updated during each move as a function of the ratio of the change in resource exception versus the change in cost.
  • the determination of whether a move is accepted is a function of the classification of the problem, and the change in the resource and cost associated with the move. For example: .. a) If the change in resource exception and the change in cost are both improvements over the previous move, then the move is accepted for all classifications; b) If the change in resource exception and the change in cost are both worse, then reject the move for all classifications; c) If the if the change in resource exception is worse, but the change in cost is an improvement, accept the move if the magnitude of the change in cost is greater than the magnitude of the change in resource exception; and d) If the change in resource exception is an improvement and the change in cost is worse, accept the move if the magnitude of the change in resource exception is greater than the magnitude of change in the cost, unless
  • the search is not in the Cost Emphasis phase and the resource exception is already better than the target trajectory;
  • the scheduling problem is Cost Constrained and the resource exception is already better than the target trajectory.
  • the scheduler may logarithmically reduce the effect of cost change as the search progresses by de-emphasizing uphill resource exception moves as the search temperature decrease, e.g., by including a loglO factor based on the number of temperature steps.
  • the second phase of the search is similar to the first phase with the exception that the normalizing component of the scaling factor remains constant. Therefore, the scaling parameter is only adjusted as the biasing component moves the resource exception closer to the target resource exception.
  • the resource scheduler 330 globally optimizes scheduling of the trains by abstracting both train movement and resources.
  • the use of the focused simulated annealing in COPES focuses attention on the critical areas.
  • the generation of move operators, although random, is more directed by allowing the constraints attached to each trip to make decisions regarding the usefulness of modifications to the global solution.
  • the use of a cost reactive scheduler may be used to develop a schedule by evaluating resource exception and the cost associated with the moves to result in a schedule which is not only resolvable, but represents a minimal cost solution.
  • the order scheduler 200 provides the schedule information to the planner/dispatcher 204, a portion of which i.e., the movement planner 202, is illustrated in greater detail in Figure 6.
  • the movement planner comprises a movement planner initializer 400, a movement planner executor 402, a physical model 404 (preferably a stand alone unit as illustrated in Figure 8), a display, a resolution options identifier 408 and a conflict resolver 410.
  • the movement planner initializer 400 receives the schedule from the order scheduler 200 of Figure 3 through the planner/dispatcher 204.
  • the movement planner initializer 400 also receives information regarding the state of the system from any suitable conventional external source, generally from the dispatching function of the planner/dispatcher 204. This information may be developed from a variety of sources such as the geolocating system (illustrated in Figure 10) or conventional track sensors for determining the location of trains in the system.
  • the schedule and the data as to the state of the railway system are used along with the definition of each of the trains and their starting point to initialize the movement planner.
  • the definition of a train may include all relevant data such as the number and type of locomotives, the number and type of cars and the weight of each of the cars.
  • the starting point of each train includes its position of the train in the system, its direction on the track, and its velocity.
  • the schedule includes: the originating point, a time of departure from the originating point and a destination point.
  • This data is a "state vector" which is supplied to the movement planner executive 402 along with a time interval which indicates the extent of time that the movement planner 202 should plan train movements.
  • the movement planner initializer 400 may be any appropriately programmed general purpose or special purpose computer.
  • the movement planner executor 402 receives the schedule and state of the systems data from the movement planner initializer 400 and is connected for two-way communications with the physical model 404 and the resolution options identifier 408. The movement planner executor 402 also receives information from the conflict resolver 410 and provides information to the planning/dispatching function through a terminal 406.
  • the movement planner executor may be any appropriately programmed general purpose or special purpose computer.
  • the movement planner executor 402 receives and records the state vector, and uses the services of the physical model 404 to advance time in increments until (a) the physical model 404 reports a train conflict, (b) a specific stop condition occurs or (c) the simulation time interval is reached.
  • the state vector at the time of the conflict is saved and the conflict is reported by the movement planner executive 402 along with the data reporting the time history of the motion of the trains.
  • the existence of and background information relating to the detected conflict is reported by the physical model 404 to the conflict resolver 410.
  • the physical model 404 follows the motion of the train once it has been provided by the movement planner executive 402 with data identifying the initial state, stopping condition and the time advanced interval.
  • the resolution options identifier 408 receives the notice of a conflict from the movement planner executor 402 and identifies the options available for the resolution thereof.
  • the conflict resolver 410 receives the identified options from the resolution options identifier 410 and performs an analysis based on the performance measure data received from terminal 332 of the order'scheduler 200, Figure 4. This evaluation is accomplished by simulating each of the options and computing the associated performance measure or figure of merit.
  • this "best” result is reported to the movement planner executor 402 for display to the dispatcher and/or the movement plan is revised to include the alternate path, if applicable, and the simulation using the physical model 404 is repeated beginning from the initial state or other recorded state.
  • Local optimization is satisfactory in a large percentage of scenarios if the schedule provided by the order scheduler 200 of Figure 3 has been sufficiently intelligent in specifying the dispatching times. If the dispatch times are not carefully specified, local optimization may lead to "lockup", he., a condition in which conflicts may no longer be resolved. Lockup occurs because the resolution of one conflict leads to or limits the alternatives for resolution of another set of conflicts.
  • Global optimization may be performed using a variety of optimization techniques, preferably a version of the well known "branch and bound” technique for searching a tree of alternative solutions.
  • each of the conflicts is modeled as a branch point on a decision tree.
  • the search technique chooses the lowest cost alternative and continues the simulation.
  • the cost of alternatives is saved, as is the state of the system for each of the conflict points. It is possible that choosing the lowest cost solution among the alternatives may not result in the optimum overall solution.
  • 61 technique allows the search to back up in the tree and retract decisions previously made in order to reach a lower cost solution or avoid a lockup.
  • the movement plan available at the dispatcher terminal 406 desirably includes a suitable conventional display to display the motion of the trains until a conflict occurs, and to present the time history leading up to the conflict in a graphical form for interpretation and resolution by a human operator.
  • the data from the optional resolution options identifier 408 may be displayed to the operator to assist him in manually resolving the conflict.
  • the conflict resolver 410 may provide a suggestion as to resolution of the conflict and that suggestion may also be displayed to the operator.
  • Constraints considered in this description generally fall into three categories, those time constraints which are inherent in the task of filling an order, those constraints which are inherent in the structure of the railroad, and those constraints which are explicitly specified by the user.
  • Order constraints include the sequential nature of the activities based upon the fact that a train cannot jump from one point to another without passing through some intermediate segments. For example, in order to load coal at a mine, a train must capture
  • Constraints are also inherent in the structure of the railroad. Such constraints include gap-able elements (sidings located between segments) and single/multiple track configurations. A wide variety of user defined constraints may be included. These constraints are generally time constraints which seek to restrict the resource scheduler 330 from scheduling certain resources over certain time periods.
  • constraints are mine which has limited hours (e.g. daylight only) during which it can load coal. Such a constraint would be included by limiting the resource availability to a specified interval.
  • resources such as track or locomotives, which are out of service for maintenance during a specified time interval.
  • train which is not under the control of the scheduler, e.g., a passenger train which is scheduled by an entity external to the freight train scheduler. All of these constraints may be included by appropriately defining the resource availability timelines.
  • the rule-based process converts orders into a form which is suited to a constraint-propagation solution and restricts the search space by eliminating certain candidate solutions, based upon a set of rules incorporating company policy, standard operating procedures and experience factors, among others.
  • the constraint-based process solves the problem of moving time intervals to maximize the externally supplied performance measure while satisfying all of the constraints.
  • the result of this process is a schedule for railway operation which includes a globally optimized schedule for train operation, maintenance activities, and terminal equipment.
  • each of the processes may be implemented as an asynchronous UNIX process with inter-process communications between the two processes implemented using a well known client server relationship based upon UNIX sockets.
  • the procedural means In the event that the procedural means is provided, it also is implemented as one or more asynchronous UNIX processes. These processes communicate using a well-known client-server inter-process communications.
  • the procedural means is used to refine the schedule to include details of the rail system. This is accomplished by simulating the operation of the railroad, identifying the conflicts in the schedule which result from the level of model abstraction used in the constraint-based process, and adjusting the schedule to eliminate those conflicts while at the same time maximizing the performance measure.
  • the movement plan obtained by refining the schedule is returned to the rule-based processor. If for any reason, all conflicts cannot be resolved, the movement plan is returned to the rule-based processor with the conflict duly noted.
  • the rule-based processor examines the movement plan based upon set of rules depicting company policies and, if the movement plan is satisfactory, forwards the movement plan to the dispatcher for display or for use in controlling the applicable trains as described infra.
  • a schedule exception is a predicted failure to meet a defined schedule which requires rescheduling of the involved
  • Extrinsic traffic is pre-scheduled traffic not to be altered by the system. Orders may arrive as a batch or arrive in a sequence over a period of time.
  • the user interface 500 translates this data into "facts” and asserts them into the rule-based process.
  • the user may also add, remove, or change certain rules in the rules database for the purpose of including company policy and other experience factors which may change over time.
  • the user interface 500 provides data to a rule based expert system 502.
  • a variety of expert system tools are available to allow the facts to be asserted and processed by a rule-based inference engine according to the rules contained in the rule data base.
  • the preferred implementation is the C-Language Integrated Production System (CLIPS) developed by NASA Johnson Space Flight Center because it is readily imbedded into a system and supports an object-oriented approach which is compatible with the constraint- based element .
  • CLIPS C-Language Integrated Production System
  • Order specific rules include rules which identify the sequence of activities with associated resources which are required to fill an order and put
  • Order specific rules also include rules which detem ine the extent to which scheduling will be performed in the event that a prior schedule exists. For example, company policy may dictate that trips scheduled to begin within a specified time period not be rescheduled upon receipt of a new order, but may be rescheduled in the event of
  • a second category is rules which receive availability information from the user interface 500 and process these rules into a form which is suitable for application to the constraint based process.
  • Availability is modified to account for extrinsic traffic, locomotives out of service for repair or maintenance, track out of service, or other factors which affect the availability profiles.
  • a third category of rules are rules which restrict the search space for the constraint based process. Rules are provided to determine the route to be taken to accomplish the order. In many of the larger railroads there are multiple paths which can be taken to move a train from one point to another. This set of rules selects the optimum path based upon principals of physics, specified performance measures, standard operating procedures or and experience factors. Trains which cannot service an order because of locomotive power or terminal equipment limitations are excluded from consideration.
  • a fourth category of rules is those rules which evaluate the schedule returned by the constraint based process and either resubmit the orders to the constraint based process after relaxing some of the constraints, submit the schedule to the procedural means (if available), or notify the user through the user interface 500 that the request is overly constrained and cannot be scheduled.
  • a fifth category of rules are those rules which evaluate the schedule and determine if it should be replanned, i.e. if there are no conflicts present, is it acceptable according to company policy and is it complete.
  • a sixth category of rules are those which receive notification of deviations of the trains from the movement plan and determine whether or not re-scheduling should occur, and if the rescheduling should be performed by adjusting the movement plan or the schedule.
  • a request to schedule an order from a scheduler client 504 may be submitted via the client-server 508 to the constraint based expert system 510 for scheduling.
  • the rule based expert system 502 determines the action to be taken. Depending upon the rules, this action may include rescheduling or, if the unresolved conflict is small, the schedule may be forwarded to the procedural means (if available) to resolve in the course of refining the schedule into the detailed movement plan.
  • the schedule may be passed to a dispatcher terminal/display 506 if desired for display to an operator (e.g. a dispatcher) or to automated dispatching. If the procedural process 516 is available, the schedule along with a performance measure may be passed there via the movement planner client 508 for refinement.
  • the scheduler client 504 may receive a schedule request from the rule-based expert system 502, translate it into a structure understood by the scheduler server 508 and submit it to the scheduler server 508.
  • This schedule request may includes one or more orders.
  • an order may contain information such as the total quantity of commodity (if the order is for bulk delivery), the earliest time that pickup can occur, the latest time for delivery, and a performance measure reflecting penalties for late delivery and/or incentives for early delivery.
  • the order may reflect the activities required to service the order and the resource types (e.g., trains) suitable for servicing the order.
  • an order may include a percent of full speed parameter and a slack time percent parameter.
  • the percent of full speed parameter indicates that the schedule should be built with the trains running at less than maximum speed, thus giving the movement planner more latitude in satisfying the resulting schedule.
  • the slack time percent provides a limited amount of cushion within which the movement planner can move the train trips to assure meeting the overall schedule.
  • the schedule client 504 receives the schedule from the constraint based system 510 via the scheduler server 508, and translate it into a fact which can be asserted in the rule-based expert system 502.
  • the schedule server 508 receives an order in the form described above and translate it into a form compatible with the constraint based expert system 510. It also translates the schedule produced by the constraint based expert system 510 into a form compatible with the scheduler client 504.
  • the scheduler server 508 and the scheduler client 504 communicate using client-server inter-process communications well known in the art.
  • the constraint propagation expert system 510 satisfies a set of constraints describing an order asserted by the scheduler server 508. All of these constraints may be included by appropriately defining the resource availability timelines.
  • Constraints specified by the user include resources, such as track or locomotives, which are out of service for maintenance and train not under the purview of the scheduler, such as an Amtrack train which is scheduled by an external entity.
  • the preferred implementation for the constraint based system 510 is the well known search technique known as simulated annealing. However, other search techniques such as genetic search may be suitable for some applications.
  • Simulated annealing may be implemented using a constraint propagation shell based upon the Waltz algorithm (described, e.g., in "Understanding Line Drawings of Scenes with Shadows," The Psychology of Computer Vision, ed. P. Winston, McGraw-Hill, New York, 1975).
  • the capability to translate the sequence of activities in the activity list to a sequence of time intervals may be provided by a commercially available train performance calculator.
  • a custom developed process based upon the Davis Equations for train motion or suitable conventional means may be used to of estimate the time required for a resource to complete a specified activity. If alternative resources are available for accomplishing an activity, then alternative intervals are defined for each activity. A list of intervals, grouped by resource and by time may thus be produced.
  • Intervals are grouped together in a logical way, typically initially on the basis of entire train trips (if applicable to a particular order).
  • a gap-able interval is an interval in a group after which a gap is allowed before the next interval in the group.
  • This representation is used to represent the presence of a siding or other capability for holding a train for an interval of time while another train passes. Capability is provided to receive the interval groups, resources available intervals, and performance measures and conduct a search for a schedule which (a) satisfies the resource availability constraints, (b) satisfies the interval constraints, and (c) minimizes the performance measures.
  • the interval groups are further subdivided or gapped, the intervals regrouped and then the
  • the resulting movement plan is forwarded to the scheduler server 508 for return to the rule-based system. If all of the constraints cannot be satisfied, the movement plan is returned along with an indication that the schedule has conflicts and the identification of the resources and activities involved in the conflict.
  • a display 506 is desirably provided to display the resulting movement plan for user examination.
  • a variety of means may be used to display the plan.
  • a popular approach is a standard stringline diagram used by railroads. As illustrated in Figure 2, the stringline is a line drawing in which the position on the track is plotted as a function of the time for each train.
  • a movement planner client 512 is provided to translate the schedule into the form of a request for planning which is compatible with the movement planner server 514.
  • the movement plan is received from the movement planner server 514 and translated into a form which is compatible with the rule based expert system 502.
  • the movement planner server 514 translates the request for movement planning into a form which is compatible with the procedural system 516.
  • the server 514 also translates the movement plan received from the procedural system 516 into a form which can be understood by the movement planner server 514.
  • the movement planner client and movement planner server 514 communicate using conventional inter-process communications.
  • the procedural system 516 receives the schedule and a state of the rail network (position of trains) from an external source and initializes a simulation capability with the definition of each of the trains and their initial point.
  • the definition of a train includes
  • the position of each train includes its position of the train, its direction on the track, and its velocity.
  • the motions of all of the scheduled trains is simulated until a train conflict occurs, a specified stop condition occurs, or the simulation time interval is reached.
  • the state vector at the time of the conflict is recorded and the options available to resolve the conflict are determined. If no conflict occurs, then the movement plan is complete and it is reported to the movement planner server 514 for forwarding to the rule based system and for execution by the planning/dispatching function.
  • the options available to resolve a conflict may be enumerated.
  • Conflicts may be classified as “meets”, “passes”, “merges”, or "crossings".
  • the options for resolution of a conflict include moving one of the trains to an alternate track to await the passing of the conflicting train.
  • the departure of a train from its origin point or other point at which it is stopped may be delayed until the way is clear.
  • Still another option is to stop one of the trains at a point along its path to allow the other train to move onto an alternate track.
  • the identification of alternate track options and options for stopping along a route are enumerated beginning with those options which are closest to the point of conflict.
  • Best performance is determined by a performance measure supplied by the rule based system. Evaluation of each option is accomplished by simulating each of the options and computing the associated performance measure. If "local optimization" is employed, the movement plan is revised to include the best alternative path (if applicable), and the
  • each of the conflicts is modelled as a branch point on a decision tree.
  • the search technique chooses the lowest cost alternative and continues the simulation.
  • the cost of alternatives may be recorded, and the state of the system may be recorded periodically. It is possible that choosing the lowest cost solution among the alternatives may not result in the optimum overall solution.
  • the branch and bound technique allows the search to back up in the tree and retract decisions previously made in order to reach a lower cost solution.
  • An important aspect of the present invention is the use of a physical model of the topology of the railway system in several levels of abstraction in the planning process.
  • the topology of a railway system may be represented with multiple levels of complexity. This provides not only the capability to model highly complex systems, but also to hide levels of complexity where such complexity is a detriment to the efficient utilization of the model.
  • an object-oriented rail topology model is composed of three fundamental elements, i.e., nodes, segments, and connectors.
  • a 72 segment is used to represent a length of rail which may be single or multiple track and is composed of an ordered collection of fragments.
  • a fragment is a piece of track which has constant grade, constant curvature, constant speed limit, and length.
  • a node may represent a complex object and may itself contain internal structure composed of nodes, segments and connectors. Connectors are used at each end of a segment to join a segment to a node, and nodes may possess an arbitrary number of connectors. Each element of the topology is provided with a unique system identifier to enable the identification of a location by reference to the system identifier.
  • a rail network is represented as a node.
  • This rail network node contains stmcture which in rum can be represented as a set of nodes connected by segments.
  • This first level of complexity models a rail network as a set of track segments connecting nodes which represent gross entities such as ports, mines, setout yards, sidings, crossovers, forks, joins, and branch points. For simple track structures such as switches and junctions, this level of detail may represent the maximum level of detail. For more complex track structures such as setout yards, further levels of complexity may be added until the entire rail network is modelled in detail.
  • the node 900 at one end of a segment may be a siding 902 or a switch 904.
  • the node 906 may represent an entire port, with multiple nodes.
  • the use of one or more nodes within a node is particularly useful in developing different degrees of abstraction in something as simple as sections of track.
  • the position of a train in a rail network is indicated by the position of the head of the train.
  • the head of the train is located by the segment identifier and an offset from the
  • the direction of the train and the length of the train may be used to locate the remainder of the train.
  • data as to the position, direction and length of a train may be used to calculate the resistance of the train, by taking into account the grade and curvature of the track fragments upon which the train is located, the train velocity and other train parameters.
  • Routing from one point to another in the system may be computed by using any network routing algorithm.
  • the well known Shortest Path First (SPF) algorithm is frequently used.
  • SPPF Shortest Path First
  • the algorithm need not use distance as the performance measure in computing path length and more complex performance measures involving grades, for example, are often useful.
  • the characteristics of the railroad rolling stock may be stored on a conventional resource database 800.
  • the type, tare weight, length, cross sectional area, loaded weight, number of axles, and streamline coefficient may be provided.
  • Unit trains are also defined in the database with an identifier, train speed limit, list of locomotive types and list of car types.
  • This resource database may be implemented in tabular form, complex data structure, or using any commercially available database.
  • the defined train objects may be propagated through the system in accordance with requests for train movement provided by the simulation manager support 802. All train movement is in accordance with the equations of physics, basic train handling principles, and well known train control rules.
  • simulation manager support may consist of an ordered list of fragments from the source to the destination of each train trip with train direction on each fragment also indicated.
  • the movement of the trains along the track is governed by simple physics equations to compute the acceleration of the train.
  • the initial acceleration of the train is bounded by the adhesion of the rails and the weight of the locomotive.
  • the acceleration of some high horsepower locomotives may be limited by the force that would cause the train to uncouple.
  • the train handling rules allow the train to accelerate with maximum acceleration subject to the available tractive force of the locomotives, maximum tractive force at the rails, and the decoupling force. These values are typically set somewhat lower than actual to allow for conservative handling of the train by an engineer. Once the scheduled speed, or speed limit (if is lower ) is attained, the tractive force of the train is set exactly equal to the resistance of the train in order to maintain the speed.
  • Train braking is applied to stop the train, to reduce speed to a lower speed limit, to avoid interfering with another train or in response to a signal, and to maintain a safe speed on a grade.
  • Many techniques are available to model train braking. The capability to anticipate braking needs is provided by searching the track ahead for speed limit changes, other trains or signals.
  • control methods are "no control”, “moving block control”, and “fixed block control”.
  • the no control method is used to move a single train through the network. The train moves through the network with no concern for the signaling system or the presence of other trains. This method is useful when computing
  • a train checks the railway signalling model at each time interval to determine if a signal is visible to the train and if so, whether the signal indicates that the train should continue, slow or stop.
  • Specific rules in the signalling system depend upon the railroad which is being modelled. The control behavior indicated by the railway signalling model supersedes all other speed limits.
  • Moving block control is based on establishing a forbidden zone associated with each train.
  • the forbidden zone for a train includes the train and a length of track in front of and along the route of the train which is equal in length to the stopping distance of the train plus any ambiguity as to the train's position.
  • the stopping distance is of course dependent upon the speed of the train, the grade, the track adhesion coefficient , and the weight of the train. This requires that each train monitor the position of the forbidden zone of other trains to assure that the forbidden zone of no other train enter its forbidden zone. To avoid such an incident, brake handling rules are applied to assure that the train decelerates in an appropriate fashion to avoid conflict.
  • the positions of the trains relative to the specified stop conditions are monitored. If a stop condition occurs, the time advance ceases and the results including a time history of the path of the trains is reported.
  • a signalling system based upon conventional fixed block signals may be modeled.
  • Signal blocks are defined and related to the fragment track structures used in the multi-level modelling of rail topology. As the head of a train occupies a fragment associated with a signalling block, the status of the block changes from "unoccupied” to "occupied”. When the tail of a train exits all fragments within a block, the block status is changed to "unoccupied”.
  • the relationship of the block status to the signals is defined by a set of company-specific railway rules which are part of a standardized set. Information on these rules may be obtained from publications of the American Association of Railroads and other sources.
  • the automatic block signalling (ABS) is well known and may be used as an illustrative implementation.
  • the signal is changed to yellow over green, corresponding to Rule 282 which requires a train to approach the next signal at restricted speed.
  • the signal changes to "clear”, and corresponding to Rule 281 which allows the train to proceed in accordance with all applicable speed limits.
  • the four level signaling system is implemented by two vertically spaced lights, i.e., red over red is stop, yellow over red is restricted speed, yellow over green is medium speed, and green over green is clear.

Landscapes

  • Engineering & Computer Science (AREA)
  • Mechanical Engineering (AREA)
  • Train Traffic Observation, Control, And Security (AREA)
  • Electric Propulsion And Braking For Vehicles (AREA)

Abstract

Système et procédé de programmation servant à déplacer plusieurs objets à travers un dispositif comportant des trajets multiples et décrit sous la forme de système de programmation de wagons de marchandises. Ce système de programmation met en application un programmateur de ressources permettant d'optimiser les coûts, de manière à limiter au maximum le recours à des ressources exceptionnelles, tout en minimisant simultanément les coûts globaux associés à cette solution. On peut utiliser cette programmation de déplacement afin de contribuer à la commande du mouvement de trains à travers le système ou de commander automatiquement ce mouvement.
PCT/US2000/035090 1999-12-31 2000-12-22 Programmateur de ressources exerçant un effet positif sur les couts et procede correspondant WO2001050666A2 (fr)

Priority Applications (2)

Application Number Priority Date Filing Date Title
AU24529/01A AU2452901A (en) 1999-12-31 2000-12-22 Cost reactive resource scheduler and method
CA002397123A CA2397123A1 (fr) 1999-12-31 2000-12-22 Programmateur de ressources exercant un effet positif sur les couts et procede correspondant

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US09/476,615 US7092894B1 (en) 1994-09-01 1999-12-31 Cost reactive scheduler and method
US09/476,615 1999-12-31

Publications (3)

Publication Number Publication Date
WO2001050666A2 true WO2001050666A2 (fr) 2001-07-12
WO2001050666A3 WO2001050666A3 (fr) 2002-03-21
WO2001050666A9 WO2001050666A9 (fr) 2002-05-30

Family

ID=23892564

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2000/035090 WO2001050666A2 (fr) 1999-12-31 2000-12-22 Programmateur de ressources exerçant un effet positif sur les couts et procede correspondant

Country Status (3)

Country Link
AU (1) AU2452901A (fr)
CA (1) CA2397123A1 (fr)
WO (1) WO2001050666A2 (fr)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1573578A2 (fr) * 2002-12-20 2005-09-14 UNION SWITCH & SIGNAL Inc. Procede et systeme d'optimisation dynamique d'une planification de trafic
CN115313487A (zh) * 2022-08-22 2022-11-08 西安交通大学 一种考虑检修流的移动氢能微网调度方法

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5265006A (en) * 1990-12-14 1993-11-23 Andersen Consulting Demand scheduled partial carrier load planning system for the transportation industry
US5467268A (en) * 1994-02-25 1995-11-14 Minnesota Mining And Manufacturing Company Method for resource assignment and scheduling
US5541848A (en) * 1994-12-15 1996-07-30 Atlantic Richfield Company Genetic method of scheduling the delivery of non-uniform inventory
US5623413A (en) * 1994-09-01 1997-04-22 Harris Corporation Scheduling system and method

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5265006A (en) * 1990-12-14 1993-11-23 Andersen Consulting Demand scheduled partial carrier load planning system for the transportation industry
US5467268A (en) * 1994-02-25 1995-11-14 Minnesota Mining And Manufacturing Company Method for resource assignment and scheduling
US5623413A (en) * 1994-09-01 1997-04-22 Harris Corporation Scheduling system and method
US5541848A (en) * 1994-12-15 1996-07-30 Atlantic Richfield Company Genetic method of scheduling the delivery of non-uniform inventory

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1573578A2 (fr) * 2002-12-20 2005-09-14 UNION SWITCH & SIGNAL Inc. Procede et systeme d'optimisation dynamique d'une planification de trafic
EP1573578A4 (fr) * 2002-12-20 2007-05-30 Union Switch & Signal Inc Procede et systeme d'optimisation dynamique d'une planification de trafic
US7386391B2 (en) 2002-12-20 2008-06-10 Union Switch & Signal, Inc. Dynamic optimizing traffic planning method and system
CN115313487A (zh) * 2022-08-22 2022-11-08 西安交通大学 一种考虑检修流的移动氢能微网调度方法

Also Published As

Publication number Publication date
AU2452901A (en) 2001-07-16
WO2001050666A9 (fr) 2002-05-30
WO2001050666A3 (fr) 2002-03-21
CA2397123A1 (fr) 2001-07-12

Similar Documents

Publication Publication Date Title
US7092894B1 (en) Cost reactive scheduler and method
US7539624B2 (en) Automatic train control system and method
AU712538B2 (en) Scheduling system and method
Goverde Punctuality of railway operations and timetable stability analysis
US8082071B2 (en) System and method of multi-generation positive train control system
US7725249B2 (en) Method and apparatus for congestion management
Kraay et al. Real-time scheduling of freight railroads
RU2470814C2 (ru) Система и способ для оптимизации работы поезда с учетом параметров вагона
US7937193B2 (en) Method and apparatus for coordinating railway line of road and yard planners
US8589057B2 (en) Method and apparatus for automatic selection of alternative routing through congested areas using congestion prediction metrics
US9896115B2 (en) System and method for coordinating terminal operations with line of road movements
Li et al. Real-time scheduling on a transit bus route: a 0-1 stochastic programming model
US7797087B2 (en) Method and apparatus for selectively disabling train location reports
WO2001050666A2 (fr) Programmateur de ressources exerçant un effet positif sur les couts et procede correspondant
Dong Modeling rail freight operations under different operating strategies
MXPA97001524A (en) System and method of programac
Arcot Modeling uncertainty in rail freight operations: Implications for service reliability
D’Ariano et al. Evaluating the performance of railway dynamic traffic management
US20070260497A1 (en) Method of planning train movement using a front end cost function
RU2481988C2 (ru) Система и способ оптимизации рейса для транспортного средства
Dong Terminal operating plan and achievability: an approach to improve rail terminal performance
D’Ariano et al. Testing the performance of railway dynamic traffic management strategies
Pranzo Real-time train conflict detection and resolution
Lewellen et al. The Rail Yard Manager

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A2

Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY BZ CA CH CN CR CU CZ DE DK DM DZ EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NO NZ PL PT RO RU SD SE SG SI SK SL TJ TM TR TT TZ UA UG US UZ VN YU ZA ZW

AL Designated countries for regional patents

Kind code of ref document: A2

Designated state(s): GH GM KE LS MW MZ SD SL SZ TZ UG ZW AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE TR BF BJ CF CG CI CM GA GN GW ML MR NE SN TD TG

121 Ep: the epo has been informed by wipo that ep was designated in this application
DFPE Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101)
AK Designated states

Kind code of ref document: A3

Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY BZ CA CH CN CR CU CZ DE DK DM DZ EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NO NZ PL PT RO RU SD SE SG SI SK SL TJ TM TR TT TZ UA UG US UZ VN YU ZA ZW

AL Designated countries for regional patents

Kind code of ref document: A3

Designated state(s): GH GM KE LS MW MZ SD SL SZ TZ UG ZW AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE TR BF BJ CF CG CI CM GA GN GW ML MR NE SN TD TG

AK Designated states

Kind code of ref document: C2

Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY BZ CA CH CN CR CU CZ DE DK DM DZ EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NO NZ PL PT RO RU SD SE SG SI SK SL TJ TM TR TT TZ UA UG US UZ VN YU ZA ZW

AL Designated countries for regional patents

Kind code of ref document: C2

Designated state(s): GH GM KE LS MW MZ SD SL SZ TZ UG ZW AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE TR BF BJ CF CG CI CM GA GN GW ML MR NE SN TD TG

COP Corrected version of pamphlet

Free format text: PAGES 2-89, DESCRIPTION, REPLACED BY NEW PAGES 2-89; PAGES 90-94, CLAIMS, REPLACED BY NEW PAGES 90-94; PAGES 1/13-13/13, DRAWINGS, REPLACED BY NEW PAGES 1/13-13/13; DUE TO LATE TRANSMITTAL BY THE RECEIVING OFFICE

WWE Wipo information: entry into national phase

Ref document number: 2397123

Country of ref document: CA

REG Reference to national code

Ref country code: DE

Ref legal event code: 8642

122 Ep: pct application non-entry in european phase
NENP Non-entry into the national phase

Ref country code: JP