Binary tree monorail multi-machine task cooperative scheduling method based on dynamic planning and real-time scheduling

Abstract

The invention provides a binary tree monorail multi-machine task cooperative scheduling method based on dynamic planning and real-time scheduling. The method comprises scheduling of a starting stage and a returning stage of a monorail multi-machine; in the starting stage, a dynamic planning strategy is adopted to gradually optimize task allocation and path planning by constructing a state transition equation and a cost function; in the course reversal stage, course reversal scheduling adjustment is carried out by dynamically adjusting a course reversal sequence and an orbital transfer strategy; and from the last-stage orbit transfer node of the binary tree orbit network, deriving from the upper-layer orbit transfer nodes step by step by taking the minimization of the total passing time of the vehicle as a target. According to the method, the global optimization capability of dynamic planning and the real-time response requirement of the scheduling task are combined, and through recursive task allocation and path optimization, not only can task change and emergency situations be quickly responded, but also the running paths of multiple machines in a complex track can be effectively coordinated, task conflicts and path congestion are avoided, and the task scheduling efficiency is improved. The optimal substructure characteristic of dynamic planning avoids the problem of high complexity of traditional global calculation.

Description

Technical Field

[0001] The present invention relates to the technical field of task scheduling, and in particular to a binary tree single-track multi-machine task collaborative scheduling method based on dynamic programming and real-time scheduling. Background Art

[0002] Transport scheduling in complex terrain, such as mountain orchards, presents numerous challenges due to densely planted fruit trees and uneven terrain. Traditional transportation methods are no longer able to meet the demands for efficient harvesting and transportation. Monorail transporters, as an emerging transportation tool, are gaining attention for their flexibility and adaptability. They can efficiently operate on narrow paths, transporting crops and equipment, and thus improving overall operational efficiency.

[0003] However, the scheduling of existing monorail transporters still remains at the more basic sequential scheduling logic. The core of this method is to arrange the monorail machines to perform tasks and return to the starting point in a fixed order, that is, the monorail machines complete their assigned tasks in a pre-set order and then return in sequence. Although this method realizes the basic function of scheduling, it has significant limitations. Since the vehicles complete their tasks independently in sequence, each monorail machine may need to wait for a long time for the vehicle in front to complete the work or pass the track change node on the way back, resulting in low track resource utilization. In addition, since the track network resources are not efficiently utilized, the vehicles will incur additional time costs during the waiting process, exacerbating the inefficiency of the overall transportation system. Moreover, this scheduling method has poor adaptability to emergencies. For example, when a vehicle is stranded due to mechanical failure or task delay, the scheduling plan of subsequent vehicles often needs to be completely reconstructed, which further increases time waste and scheduling complexity.

[0004] To improve scheduling efficiency, existing technologies attempt to employ algorithms for scheduling improvements. For example, through task sharing and optimization algorithms, tasks are distributed as evenly as possible across monorail vehicles to avoid overloading or idleness of some vehicles. However, these methods often use fixed vehicle task paths, optimizing only task allocation and exhibiting limited adaptability to dynamic task changes. Furthermore, in special circumstances, such as complex track network structures, existing solutions are unable to resolve track change conflicts or real-time dynamic control of multiple vehicles, failing to provide a systematic solution. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, the present invention provides a binary tree single-track multi-machine task collaborative scheduling method based on dynamic programming and real-time scheduling. The present invention adopts a dynamic programming algorithm to achieve global optimal path planning through recursive task allocation in the departure phase; in the return phase, through bottom-up node priority determination, multi-vehicle conflicts are coordinated in real time. Combined with the "optimal substructure" feature of dynamic programming, the present invention further reduces computational complexity and improves response speed.

[0006] The technical solution of the present invention is: a binary tree single-track multi-machine task collaborative scheduling method based on dynamic programming and real-time scheduling, the scheduling method comprising a single-track multi-machine departure phase and a return phase; the single-track multi-machine departure phase adopts a dynamic programming strategy to gradually optimize task allocation and path planning by constructing a state transition equation and a cost function; in the return phase, return scheduling is adjusted by dynamically adjusting the return sequence and track change strategy; starting from the last-level track change node of the binary tree track network, with the goal of minimizing the total vehicle travel time, the method is gradually derived towards the upper-level track change nodes.

[0007] Preferably, the binary tree track network uses the monorail starting point as the root node and the track change structure as the branch node and leaf node; and all nodes are connected by tracks.

[0008] Preferably, the single-track multi-machine departure stage specifically includes the following steps:

[0009] S1), initialize the priority of the task point and the initial state of the monorail machine;

[0010] S2) Allocate task points to the monorail machine with the highest priority according to the total cost of the task points and the task weight;

[0011] S3), repeat step S2) for the remaining monorail machines t2, t3..., t m Assign task points one by one until all task points are assigned or the monorail queue is empty.

[0012] Preferably, the scheduling of the return phase of the monorail machine includes the following steps:

[0013] A1) Initialize the position of the monorail machine and the status of the track change node;

[0014] A2) Determine the scheduling priority of a single node based on the total cost of the single-track machine;

[0015] A3) Update the status of the monorail machine and nodes;

[0016] A4) Using the scheduling result of step A3) as the input of the upper layer node to recursively update the scheduling state;

[0017] A5) Repeat the priority determination and status update of steps A2) and A3) until the root node is reached.

[0018] Preferably, in step S1), initializing the priority of the task point and the initial state of the monorail machine specifically includes the following steps:

[0019] S11) Construct a set of task points M, namely:

[0020] M={m1,m2,…,m n}; (1)

[0021] Among them, m i represents the i-th task point;

[0022] S12) Record the position pos of each monorail machine j , and calculate the position pos of each monorail machine j The shortest distance d from the starting point root j ,Right now:

[0023] d j =dist(pos j ,root); (2)

[0024] Where dist represents the shortest path length function in the binary tree track network;

[0025] S13) All the monorail machines are arranged according to the distance d between them and the starting point root. j Sort the single-track machine t in ascending order to obtain the priority queue Queue of the single-track machine, that is:

[0026] Queue={t1,t2,…,t m}, d1≤d2≤…≤d m ; (3)

[0027] Among them, the monorail machine close to the starting point root has priority in task allocation.

[0028] Preferably, in step S2), the task points are assigned to the monorail machine with the highest priority, specifically comprising the following steps:

[0029] S21) Select the monorail machine t1 with the highest priority from the monorail machine priority queue Queue, and calculate the total cost C of traveling to each task point based on its current position pos1 and the task point set M. 1j :

[0030]

[0031] Where, dist(pos1,m j ) indicates that the monorail machine t1 moves from its position pos1 to the task point m j The shortest distance; represents the track switching cost, r represents the number of switching points; v l Indicates the lth switching point; switch indicates the switching operation;

[0032] S22) From all task points, combined with task weight W jand distance D j Select a task point m that meets the following conditions j ,Right now:

[0033]

[0034] In the formula, α and β are weight parameters, emphasizing the priority of high-priority task points and the farthest task points; D j =dist(root,m j ) represents the starting point root and the task point m j The shortest distance; αW j Indicates that high-weight tasks are given priority. When α>0, -αW j The smaller the βD, the smaller the overall cost; j Indicates that distant tasks are assigned first according to the weight coefficient β to avoid blocking subsequent vehicles with tasks in the middle;

[0035] S23) The task point m selected in step S22) is j Assign it to the monorail machine t1, and update the task point set M and the position of the monorail machine t1, that is:

[0036] M=M-{m j}; (6)

[0037] pos1=m j (7)

[0038] As a preference, in step S3), before each assignment, the current task point set M and the updated monorail machine position pos need to be considered. i ; and dynamically calculate each monorail machine t i Go to each mission point m k The total cost C ik ; And after each task point is assigned, the state cost function F(S k ) to measure the path cost at the current stage, namely:

[0039]

[0040] Where S k represents the state set after the k-th step scheduling; F(S k ) indicates reaching state S k The minimum cumulative path cost when .

[0041] As a preference, in step S3), and after each task point is assigned, calculate each monorail machine t i The path efficiency η i ,Right now:

[0042]

[0043] Where W k Point m for the task k The task weight, C ik For monorail machine i Go to mission point m k The total cost of n i Indicates monorail machine t i The number of mission points assigned.

[0044] Preferably, in step A1), the current task completion time and position of each monorail machine are recorded, and an initial state set S is constructed, namely:

[0045] S={(C1,T1),(C2,T2),…,(C n ,T n )}; (10)

[0046] Where C i Indicates monorail machine t i Current node; T i Indicates monorail machine t i Time to reach the node on the previous layer;

[0047] For each track change node N j , record the initial state of each track change node, namely:

[0048] The current track change direction is: D(N j )∈{left, right};

[0049] The time window in which the node has been used: T used (N j ).

[0050] Preferably, in step A2), determining the single node scheduling priority specifically includes the following steps:

[0051] A21) Assume that the track change node N j There are monorails on the left and right branches that are about to arrive. Calculate the earliest arrival time of the monorails on the left and right branches, which is:

[0052] T 左到达 =min{T k ∣C k ∈Left branch vehicle}; (11-1)

[0053] T 右到达 =min{T l ∣C l ∈Right branch vehicle}; (11-2)

[0054] Where, T 左到达 、T 右到达It indicates the time when the left and right branch monorail machines arrive at the track change node; T k ∣C k Indicates that under the scheduling condition C k Under this condition, the earliest time T that a monorail machine can complete a task or pass a node is k ; min{·} means selecting the one with the smallest total cost or total time from multiple possible scheduling strategies or travel paths;

[0055] A22) Set the track change operation time to T 变 The time it takes for a single-track machine to pass through the track change node is T 过 ; Calculate the total cost of left and right track changes, and choose the single-track machine with the smallest total cost to pass first, that is:

[0056] T 右优先 =T 右到达 +T 过 +max(T 右到达 +T 过 +T 变 , T 左到达 ); (12-1)

[0057] T 左优先 =T 左到达 +T 过 +max(T 左到达 +T 过 +T 变 , T 右到达 ); (12-2)

[0058] Where, T 右优先 、T 左优先 Respectively represent the total cost of the right and left monorail machines to pass first; max() means taking the maximum value; max(T 右到达 +T 过 +T 变 , T 左到达 ) represents the time required for the left monorail machine to pass after waiting; max(T 左到达 +T 过 +T 变 , T 右到达 ) indicates the time required for the monorail machine on the right to pass after waiting;

[0059]

[0060] If the current track change direction is the right single track machine (D(N j )=right), the monorail machine on the left needs to wait for the track change to be completed; vice versa.

[0061] Preferably, in step A3), after determining the priority direction of the current node, the monorail machine and the node status are synchronously updated; the steps include:

[0062] A31) Update the time for priority access to the monorail:

[0063] Assume that the monorail machine on the right has priority and update its passing time T 右通过 :

[0064] T 右通过 =T 右到达 +T 过 ; (14)

[0065] A32) Update the time for waiting for the monorail machine to pass:

[0066] Assuming that the left monorail machine needs to wait for the track change to complete before passing, the time T for the left monorail machine to pass is 左通过 for:

[0067] T 左通过 =max(T 右通过 +T 变 , T 左到达 )+T 过 ; (15)

[0068] A33) Recording node usage time window

[0069] Current node N j Total usage time:

[0070] T 节点总时间 =max(T 右通过 ,T 左通过 ); (16)

[0071] Where, T 节点总时间 is the current node N j Total usage time.

[0072] Preferably, in step A5), after the scheduling of the root node is completed, the total time of the global scheduling is the final usage time of the root node.

[0073] The beneficial effects of the present invention are:

[0074] 1. This invention combines the global optimization capabilities of dynamic programming with the real-time response requirements of scheduling tasks. Through recursive task allocation and path optimization, it can not only quickly respond to task changes and emergencies, but also effectively coordinate the operation paths of multiple machines on complex tracks to avoid task conflicts and path congestion. At the same time, the optimal substructure characteristics of dynamic programming avoid the high complexity of traditional global calculations.

[0075] 2. The present invention divides task scheduling into a departure phase and a return phase. In the departure phase, a dynamic programming method is used for step-by-step optimization, achieving global optimization of task allocation and minimizing track conflicts and resource waste through dynamic adjustment. In the return phase, overall efficiency is optimized by dynamically adjusting the return sequence and track change strategy, avoiding path congestion and waiting time, and further improving the adaptability of task scheduling.

[0076] 3. The present invention supports multi-path parallel operation and dynamic track change through the branching design of the binary tree network, significantly improving the coverage capability under complex terrain. The monorail transport aircraft can flexibly switch paths according to mission requirements, avoiding waste of track resources and adapting to sudden mission changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 This is a schematic diagram of the structure of a binary tree track network according to an embodiment of the present invention;

[0078] Figure 2 This is a schematic diagram of the process of scheduling a monorail vehicle during the departure phase according to an embodiment of the present invention;

[0079] Figure 3 The figure is a flow chart of the scheduling of the return phase of a monorail machine according to an embodiment of the present invention. DETAILED DESCRIPTION

[0080] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings:

[0081] like Figure 1-3 As shown, this embodiment provides a binary tree single-track multi-machine task collaborative scheduling method based on dynamic programming and real-time scheduling, the scheduling method includes a single-track multi-machine departure phase and a return phase; the single-track multi-machine departure phase adopts a dynamic programming strategy to gradually optimize task allocation and path planning by constructing a state transition equation and a cost function; in the return phase, the return scheduling is adjusted by dynamically adjusting the return sequence and track change strategy; starting from the last level track change node of the binary tree track network, with the goal of minimizing the total vehicle travel time, the method is gradually derived to the upper level track change nodes.

[0082] As preferred in this embodiment, Figure 1 As shown, the binary tree track network takes the starting point of the single-track machine as the root node and the track change structure as the branch node. The branch node is also connected to multiple leaf nodes; and all nodes are connected through tracks.

[0083] In this embodiment, the root node serves as the central hub of the binary tree rail network, is the departure and return point for all monorail machines, and usually serves as the loading and unloading center of the transportation network; monorail machines load operating equipment or unload transported goods here. The number of monorail machines docked here depends on the layout length of the root node and the complexity of the binary tree transportation network.

[0084] As the most critical track intersection in the binary tree track network, the branch node is responsible for selecting the monorail machine operation path and allocating tasks;

[0085] All monorail machines must depart from the root node and travel to different mission points, returning to the starting point or a designated location upon completion. The scheduling process involves task allocation, path planning, track switching, and other complex dynamic scenarios, including conflicting paths among multiple monorail machines and track change congestion.

[0086] As preferred in this embodiment, Figure 2 As shown, the departure stage of the single-track multi-machine system specifically includes the following steps:

[0087] S1) Initialize the priority of the task point and the initial state of the monorail machine; specifically, the steps include:

[0088] S11) Construct a set of task points M, namely:

[0089] M={m1,m2,…,m n}; (1)

[0090] Among them, m i Represents the i-th task point, each task point has a corresponding task weight W j , indicating the urgency or importance of the task;

[0091] S12) Record the position pos of each monorail machine j , and calculate the position pos of each monorail machine j The shortest distance d from the starting point root j ,Right now:

[0092] d j =dist(pos j ,root); (2)

[0093] Where dist represents the shortest path length function in the binary tree track network;

[0094] S13) All the monorail machines are arranged according to the distance d between them and the starting point root. j Sort the single-track machine t in ascending order to obtain the priority queue Queue of the single-track machine, that is:

[0095] Queue={t1,t2,…,t m}, d1≤d2≤…≤d m ; (3)

[0096] Among them, monorail machines close to the starting point root are given priority in task allocation, thereby reducing path conflicts and resource waste. At the same time, combined with the priority rule, high-priority monorail machines are dispatched first to complete the tasks at the farthest task point, avoiding them stopping at intermediate task points and blocking subsequent vehicles.

[0097] S2) Allocate task points to the monorail machine with the highest priority according to the total cost and task weight of the task points; specifically, the steps include:

[0098] S21) Select the monorail machine t1 with the highest priority from the priority queue of the monorail machine, and calculate its route to each task point m according to its current position pos1 and the task point set M. j The total cost C 1j :

[0099]

[0100] Where, dist(pos1,m j ) indicates that the monorail machine t1 moves from its position pos1 to the task point m j The shortest distance; represents the track switching cost, r represents the number of switching points; v l Indicates the lth switching point; switch indicates the switching operation;

[0101] S22) From all task points, combined with task weight W j and distance D j Select a task point m that meets the following conditions j ,Right now:

[0102]

[0103] In the formula, α and β are weight parameters, emphasizing the priority of high-priority task points and the farthest task points; D j =dist(root,m j ) represents the starting point root and the task point m j The shortest distance; αW j Indicates that high-weight tasks are given priority. When α>0, -αW j The smaller the βD, the smaller the overall cost; j Indicates that distant tasks are assigned first according to the weight coefficient β to avoid blocking subsequent vehicles with tasks in the middle;

[0104] S23) The task point m selected in step S22) is j Assign it to the monorail machine t1, and update the task point set M and the position of the monorail machine t1, that is:

[0105] M=M-{m j}; (6)

[0106] pos1=m j (7)

[0107] S3), repeat step S2) for the remaining monorail machines t2, t3..., t m Assign task points one by one until all task points are assigned or the monorail queue is empty;

[0108] In this embodiment, before each allocation, the current task point set M and the updated monorail machine position pos need to be considered. i ; and dynamically calculate each monorail machine t i Go to each mission point m k The total cost C ik ; And after each task point is assigned, the state cost function F(S k ) to measure the path cost at the current stage, namely:

[0109]

[0110] Where S k represents the state set after the k-th step scheduling; F(S k ) indicates reaching state S k The minimum cumulative path cost when .

[0111] And after each task point is assigned, calculate each monorail machine t i The path efficiency η i ,Right now:

[0112]

[0113] Where W k Point m for the task k The task weight, C ik For monorail machine i Go to mission point m k The total cost of n i Indicates monorail machine t i The number of mission points assigned.

[0114] As preferred in this embodiment, when multiple monorail machines are ready to return to the starting point after completing their tasks, the scheduling of the track change nodes becomes a key link affecting global efficiency. Since the monorail machines may arrive at the same track change node at different times, or need to return to the starting point through the same track segment, it is difficult to avoid resource conflicts and time waste by relying solely on local optimization scheduling strategies, thereby reducing overall operating efficiency. Therefore, this embodiment needs to handle the dynamic arrival of multiple vehicles, the parallel operation of multiple tasks, and the timing conflicts of track change operations. This embodiment is based on a dynamic programming algorithm based on layer-by-layer recursion and global optimization as a scheduling strategy. By layer-by-layer recursion, starting from the last level track change node of the binary tree track network, with the goal of minimizing the total vehicle travel time, the upper level track change nodes are gradually derived, and the global optimal scheduling scheme of the entire binary tree is finally obtained. This embodiment can effectively handle scheduling conflicts between nodes, while using the results of recursion to shorten the calculation time and meet the real-time scheduling requirements of the track network. This embodiment abstracts each track change node into an independent binary selection model, and gradually constructs a global optimal solution based on the local optimum. By comprehensively considering the arrival time, passing time and track change operation time of the left and right branch monorail machines, the optimal track change direction, monorail machine passage sequence and estimated passage time are given at each level node, such as Figure 3 As shown, the return phase includes the following steps:

[0115] A1) Initialize the position of the monorail machine and the status of the track change node;

[0116] Record the current task completion time and position of each monorail machine and construct the initial state set S, namely:

[0117] S={(C1,T1),(C2,T2),…,(C n ,T n )}; (10)

[0118] Where C i Indicates monorail machine t i Current node; T i Indicates monorail machine t i Time to reach the node on the previous layer;

[0119] For each track change node N j , record the initial state of each track change node, namely:

[0120] The current track change direction is: D(N j )∈{left, right};

[0121] The time window in which the node has been used: T used (N j ).

[0122] A2) Determine the scheduling priority of a single node based on the total cost of the monorail machine; specifically, the steps include:

[0123] A21) Assume that the track change node N j There are monorails on the left and right branches that are about to arrive. Calculate the earliest arrival time of the monorails on the left and right branches, which is:

[0124] T 左到达 =min{T k ∣C k ∈Left branch vehicle}; (11-1)

[0125] T 右到达 =min{T l ∣C l ∈Right branch vehicle}; (11-2)

[0126] Where, T 左到达 、T 右到达 It indicates the time when the left and right branch monorail machines arrive at the track change node; T k ∣C k Indicates that under the scheduling condition C k Under this condition, the earliest time T that a monorail machine can complete a task or pass a node is k ; min{·} means selecting the one with the smallest total cost or total time from multiple possible scheduling strategies or travel paths;

[0127] A22) Set the track change operation time to T 变 The time it takes for a single-track machine to pass through the track change node is T 过 ; Calculate the total cost of left and right track changes, and choose the single-track machine with the smallest total cost to pass first, that is:

[0128] T 右优先 =T 右到达 +T 过 +max(T 右到达 +T 过 +T 变 , T 左到达 ); (12-1)

[0129] T 左优先 =T 左到达 +T 过 +max(T 左到达 +T 过 +T 变 , T 右到达 ); (12-2)

[0130] Where, T 右优先 、T 左优先 Respectively represent the total cost of the right and left monorail machines to pass first; max(T 右到达 +T过 +T 变 , T 左到达 ) represents the time required for the left monorail machine to pass after waiting; max(T 左到达 +T 过 +T 变 , T 右到达 ) indicates the time required for the monorail machine on the right to pass after waiting;

[0131]

[0132] If the current track change direction is the right single track machine (D(N j )=right), the monorail machine on the left needs to wait for the track change to be completed; vice versa.

[0133] A3) Update the status of the monorail machine and nodes;

[0134] After determining the priority direction of the current node, the monorail machine and node status are updated synchronously; the following steps are included:

[0135] A31) Update the time for priority access to the monorail:

[0136] Assume that the monorail machine on the right has priority and update its passing time T 右通过 :

[0137] T 右通过 =T 右到达 +T 过 ; (14)

[0138] A32) Update the time for waiting for the monorail machine to pass:

[0139] Assuming that the left monorail machine needs to wait for the track change to complete before passing, the time T for the left monorail machine to pass is 左通过 for:

[0140] T 左通过 =max(T 右通过 +T 变 , T 左到达 )+T 过 ; (15)

[0141] A33) Recording node usage time window

[0142] Current node N j Total usage time:

[0143] T 节点总时间 =max(T 右通过 ,T 左通过 ); (16)

[0144] Where, T 节点总时间 is the current node Nj Total usage time.

[0145] A4) Using the scheduling result of step A3) as the input of the upper layer node to recursively update the scheduling state;

[0146] T 左输入 =T 节点总时间 (left child); (17-1)

[0147] T 右输入 =T 节点总时间 (right child node); (17-2)

[0148] Where, T 左输入 、T 右输入 Respectively represent the time value of the left child node direction input of the left (right) track change node; T 节点总时间 (left child), T 节点总时间 (Right child node) represents the total occupied time from activation to release of the left (right) side track change node when the single-track machine has priority passage.

[0149] A5) Repeat the priority determination and status update of step A2) and step A3) until reaching the root node;

[0150] In this embodiment, after the scheduling of the root node is completed, the total time of global scheduling is the final usage time of the root node:

[0151] T 全局 =max(T 根左输入 ,T 根右输入 ); (18)

[0152] Where, T 根左输入 、T 根右输入 are the usage time of the left and right branches of the root node respectively; max(T 根左输入 ,T 根右输入 ) means taking T 根左输入 ,T 根右输入 The maximum value in .

[0153] The above embodiments and descriptions are only for explaining the principles and best embodiments of the present invention. Without departing from the spirit and scope of the present invention, the present invention may be subject to various changes and improvements, which shall fall within the scope of the invention to be protected.

Claims

1. A binary tree single-track multi-machine task collaborative scheduling method based on dynamic programming and real-time scheduling, characterized by: The scheduling method includes scheduling of the departure phase and return phase of a single-track multi-machine system; the departure phase adopts a dynamic programming strategy to gradually optimize task allocation and path planning by constructing a state transfer equation and a cost function; During the return phase, the return scheduling is adjusted by dynamically adjusting the return sequence and track change strategy; starting from the last level track change node of the binary tree track network, with the goal of minimizing the total vehicle travel time, it is gradually derived to the upper level track change nodes.

2. The binary tree single-track multi-machine task collaborative scheduling method based on dynamic programming and real-time scheduling according to claim 1 is characterized by: The scheduling of the departure phase of the single-track multi-machine system specifically includes the following steps: S1), initialize the priority of the task point and the initial state of the monorail machine; S2) Allocate task points to the monorail machine with the highest priority according to the total cost of the task points and the task weight; S3) Repeat step S2) to assign task points to the remaining monorail machines one by one until all task points are assigned or the monorail machine queue is empty.

3. The binary tree single-track multi-machine task collaborative scheduling method based on dynamic programming and real-time scheduling according to claim 1 is characterized by: The scheduling of the return phase of the monorail machine includes the following steps: A1) Initialize the position of the monorail machine and the status of the track change node; A2) Determine the scheduling priority of a single node based on the total cost of the single-track machine; A3) Update the status of the monorail machine and nodes; A4) Using the scheduling result of step A3) as the input of the upper layer node to recursively update the scheduling state; A5) Repeat the priority determination and status update of step A2) and step A3) until reaching the root node.

4. The binary tree single-track multi-machine task collaborative scheduling method based on dynamic programming and real-time scheduling according to claim 2 is characterized by: In step S1), the priority of the task point and the initial state of the monorail machine are initialized, which specifically includes the following steps: S11) Construct a set of task points M, namely: M={m1,m2,…,m n }; (1) Among them, m i represents the i-th task point; S12) Record the position pos of each monorail machine j , and calculate the position pos of each monorail machine j The shortest distance d from the starting point root j ,Right now: d j =dist(pos j ,root); (2) Where dist represents the shortest path length function in the binary tree track network; S13) All the monorail machines are arranged according to the distance d between them and the starting point root. j Sort the single-track machine t in ascending order to obtain the priority queue Queue of the single-track machine, that is: Queue={t1,t2,…,t m }, d1≤d2≤…≤d m ; (3) Among them, the monorail machine close to the starting point root has priority in task allocation.

5. The binary tree single-track multi-machine task collaborative scheduling method based on dynamic programming and real-time scheduling according to claim 4 is characterized by: In step S2), the task points are assigned to the monorail machine with the highest priority, which specifically includes the following steps: S21) Select the monorail machine t1 with the highest priority from the monorail machine priority queue Queue, and calculate the total cost C of traveling to each task point based on its current position pos1 and the task point set M. 1j : Where, dist(pos1,m j ) indicates that the monorail machine t1 moves from its position pos1 to the task point m j The shortest distance; represents the track switching cost, r represents the number of switching points; v l Indicates the lth switching point; switch indicates the switching operation; S22) From all task points, combined with task weight W j and distance D j Select a task point m that meets the following conditions j ,Right now: In the formula, α and β are weight parameters, emphasizing the priority of high-priority task points and the farthest task points; D j =dist(root,m j ) represents the starting point root and the task point m j The shortest distance; αW j Indicates that high-weight tasks are given priority; βD j Indicates that distant task points are allocated preferentially according to the weight coefficient β; S23) The task point m selected in step S22) is j Assign it to the monorail machine t1, and update the task point set M and the position of the monorail machine t1, that is: M=M-{m j }; (6) pos1=m j 。(7)。 6. The binary tree single-track multi-machine task collaborative scheduling method based on dynamic programming and real-time scheduling according to claim 5 is characterized by: In step S3), before each assignment, the current task point set M and the updated monorail machine position pos need to be considered. i ; and dynamically calculate each monorail machine t i Go to each mission point m k The total cost C ik ; And after each task point is assigned, the state cost function F(S k ) to measure the path cost at the current stage, namely: Where S k represents the state set after the k-th step scheduling; F(S k ) indicates reaching state S k The minimum cumulative path cost when .

7. The binary tree single-track multi-machine task collaborative scheduling method based on dynamic programming and real-time scheduling according to claim 6 is characterized by: In step S3), after each task point is assigned, calculate the i The path efficiency η i ,Right now: Where W k Point m for the task k The task weight, C ik For monorail machine i Go to mission point m k The total cost of n i Indicates monorail machine t i The number of mission points assigned.

8. The binary tree single-track multi-machine task collaborative scheduling method based on dynamic programming and real-time scheduling according to claim 3 is characterized by: In step A1), the current task completion time and position of each monorail machine are recorded, and the initial state set S is constructed, namely: S={(C1,T1),(C2,T2),…,(C n ,T n )}; (10) Where C i Indicates monorail machine t i Current node; T i Indicates monorail machine t i Time to reach the node on the previous layer; For each track change node N j , record the initial state of each track change node, namely: The current track change direction is: D(N j )∈{left, right}; The time window in which the node has been used: T used (N j ).

9. The binary tree single-track multi-machine task collaborative scheduling method based on dynamic programming and real-time scheduling according to claim 8, characterized in that: In step A2), the single node scheduling priority is determined, specifically including the following steps: A21) Assume that the track change node N j There are monorails on the left and right branches that are about to arrive. Calculate the earliest arrival time of the monorails on the left and right branches, which is: T 左到达 =min{T k ∣C k ∈Left branch vehicle}; (11-1) T 右到达 =min{T l ∣C l ∈Right branch vehicle}; (11-2) Where, T 左到达 、T 右到达 Indicates the time when the left and right branch monorail machines arrive at the track change node; T k ∣C k Indicates that under the scheduling condition C k Next, monorail machine k The earliest time T to complete the task or pass the node k ; min{·} means selecting the one with the smallest total cost or total time from multiple possible scheduling strategies or travel paths; A22) Set the track change operation time to T 变 The time it takes for a single-track machine to pass through the track change node is T 过 ; Calculate the total cost of left and right track changes, and choose the single-track machine with the smallest total cost to pass first, that is: T 右优先 =T 右到达 +T 过 +max(T 右到达 +T 过 +T 变 ,T 左到达 ); (12-1) T 左优先 =T 左到达 +T 过 +max(T 左到达 +T 过 +T 变 ,T 右到达 ); (12-2) Where, T 右优先 、T 左优先 Respectively represent the total cost of the right and left monorail machines to pass first; max() means taking the maximum value; max(T 右到达 +T 过 +T 变 , T 左到达 ) represents the time required for the left monorail machine to pass after waiting; max(T 左到达 +T 过 +T 变 , T 右到达 ) indicates the time required for the monorail machine on the right to pass after waiting; If the current track change direction is the right single track machine (D(N j )=right), the monorail machine on the left needs to wait for the track change to be completed; vice versa.

10. The binary tree single-track multi-machine task collaborative scheduling method based on dynamic programming and real-time scheduling according to claim 9, characterized in that: In step A3), after determining the priority direction of the current node, the monorail machine and the node status are updated synchronously; the steps include: A31) Update the time for priority access to the monorail: Assume that the monorail machine on the right has priority and update its passing time T 右通过 : T 右通过 =T 右到达 +T 过 ; (14) A32) Update the time for waiting for the monorail machine to pass: Assuming that the left monorail machine needs to wait for the track change to complete before passing, the time T for the left monorail machine to pass is 左通过 for: T 左通过 =max(T 右通过 +T 变 ,T 左到达 )+T 过 ; (15) A33) Recording node usage time window Current node N j Total usage time: T 节点总时间 =max(T 右通过 ,T 左通过 ); (16) Where, T 节点总时间 is the current node N j Total usage time.

Images