Integrated coal supply chain scheduling decision-making method and system

Abstract

The invention discloses an integrated coal supply chain scheduling decision-making method and system, and the method comprises the steps: constructing a plurality of basic scheduling optimization algorithms covering the railway transportation of a coal supply chain to the port shipping, comprising a coal train loading optimization algorithm, a coal train delivery backspace optimization algorithm, a railway downstream heavy train shunting optimization algorithm, a port ship arrangement and shipping plan optimization algorithm and a ship replacement optimization algorithm. And according to a typical scene in a scheduling process, carrying out combined calling on the basic optimization algorithm, and constructing a decision optimization model in the typical scene. According to the method, multiple scheduling optimization algorithms are automatically combined and called for multiple scenes such as insufficient loading, transport capacity reduction, port abnormity and power plant inventory emergency, rapid response and stable operation of the whole coal transportation supply chain are achieved, and the overall scheduling efficiency of the supply chain is remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of coal transportation, and in particular relates to an integrated coal supply chain scheduling decision-making method and system. Background Technology

[0002] my country's coal resources are mainly distributed in the northwest region, while coal consumption is mainly concentrated in the eastern and southern coastal areas. Integrated coal transportation has become the main operating mode for large domestic energy companies.

[0003] An integrated coal supply chain involves multiple stages, including production, transportation, sales, storage, and consumption. It encompasses various transportation methods such as rail transport, port handling, and ship dispatch. An anomaly in any stage can disrupt the entire supply chain, impacting upstream coal production and transportation, and downstream power plant coal supply security. Large energy groups often experience low coal dispatch efficiency or even supply chain disruptions during integrated coal supply chain scheduling due to external factors or flawed internal planning, resulting in significant economic losses.

[0004] The integrated coal supply chain of large energy groups is a typical complex system with large scale, complex structure, many influencing factors, and rapid dynamic changes. Existing integrated coal dispatching optimization methods are mostly geared towards planning and dispatching on long-term scales such as annual and monthly, using coal supply chain mechanism modeling to solve the economic problems of annual or monthly planning, or optimizing coal transportation only for a single link. For example, patent CN115660171A provides a logistics network balancing optimization method that considers the coordination of the entire chain of production, transportation, sales, storage, and use. The method optimizes and adjusts the annual total plan for coal production, transportation, and sales by constructing a linear programming model with the goal of maximizing coal profits and constraints such as railway transport capacity and upper and lower limits on coal production and sales. Chinese patent CN116227804A provides a method and device for optimizing the scheduling of heavy coal trains in railway transportation based on spatiotemporal networks. It uses spatiotemporal networks to construct a downstream railway transportation scheduling model and adjusts the operation strategy of heavy coal trains to ensure the smooth operation of downstream trains. Patent CN116957440A provides a coal transportation planning method and system for large power generation groups. The method mainly focuses on power plant enterprises using linear programming to optimize the ship route plan for coal procurement plans of hydropower plants, thereby maximizing power plant procurement and coal planning.

[0005] In summary, existing integrated coal dispatching optimization methods are mostly geared towards a specific link in the integrated coal dispatching process. For example, Chinese patent CN116227804A mainly focuses on the loaded train operation and diversion of downstream railways, while patent CN116957440A mainly focuses on the optimization of coal transportation and procurement for downstream hydropower plants. Alternatively, they optimize the integrated coal production, transportation, and sales plan from the perspective of annual planning, rather than focusing on specific integrated daily coal dispatching. The existing technologies as a whole lack consideration for the overall dispatching perspective of large energy groups, supporting practical integrated dispatching strategies for coal production and loading, railway train loading and unloading, downstream loaded train operation and diversion, port vessel dispatching, and coal use by downstream hydropower plants. Summary of the Invention

[0006] To address the shortcomings of existing technologies, such as the lack of a comprehensive solution for integrated coal supply chain scheduling and decision optimization from the perspective of overall energy group scheduling, and the lack of a comprehensive solution for abnormal scheduling scenarios in the integrated supply chain, this invention provides an integrated coal supply chain scheduling decision-making method and system. The method includes: constructing multiple basic scheduling optimization algorithms covering coal supply chain transportation from railway to port shipping, including coal train loading optimization algorithms, coal train load-and-load return optimization algorithms, railway downstream loaded car diversion optimization algorithms, port vessel scheduling and shipping plan optimization algorithms, and vessel replacement optimization algorithms; and combining and calling these basic optimization algorithms according to typical scenarios in the scheduling process to construct decision optimization models for typical scenarios. This invention automatically combines and calls multiple scheduling optimization algorithms for multiple scenarios such as insufficient loading, reduced transport capacity, port anomalies, and power plant inventory shortages, achieving rapid response and stable operation across the entire coal transportation supply chain, significantly improving the overall scheduling efficiency of the supply chain.

[0007] The present invention adopts the following technical solution.

[0008] The first aspect of the present invention provides an integrated coal supply chain scheduling decision-making method, comprising: Construct a variety of basic scheduling optimization algorithms covering the coal supply chain from railway transportation to port shipping; For different scenarios, a combination of basic scheduling optimization algorithms is used; when the expected loading volume of upstream loading stations is lower than the original loading plan, a decision model for insufficient coal loading is constructed; by calculating the number of coal trains leaving the area, the deviation of the number of loading trains at loading stations, and the weight of loaded cars, the algorithms for coal train loading optimization, loaded car return optimization, and downstream heavy car diversion optimization are combined and called. Based on the real-time transport capacity of each railway line, the transport capacity of the railway line on that day is calculated; when the transport capacity on that day is less than the planned transport capacity, a decision model for the scenario of reduced transport capacity in the railway section is constructed; by calculating the transport capacity and the weight of loaded cars on the railway route, the optimization algorithms for coal trains returning empty after being loaded, loading optimization, and downstream loaded car diversion optimization are combined and called. When the unloading capacity of a station is lower than the predetermined unloading capacity, a scenario model of reduced heavy car handling capacity at key stations is constructed; by calculating the transport capacity and heavy car delivery weight of the railway route during railway maintenance, the optimization algorithms for coal trains returning empty after heavy delivery, loading optimization, and heavy car diversion optimization of the downstream railway are combined and called. When the port's unloading capacity is lower than the predetermined capacity, a decision-making model for abnormal port unloading scenarios is constructed. By calculating the port's unloading volume during maintenance and the actual number of trains diverted for heavy traffic, the algorithms for optimizing the diversion of heavy trains downstream of the railway, optimizing the return of coal trains to empty tracks, and optimizing loading are combined and called. When a port prohibits ships from entering or leaving, a decision-making model for a port maritime closure scenario is constructed. By calculating the port inventory and the actual number of trains diverted and loaded, the optimization algorithms for downstream heavy train diversion, coal trains returning empty after heavy loading and loading are combined and called. When a ship malfunction causes the corresponding power plant to be unable to receive coal on time or its inventory to fall below the predetermined safe inventory level, a decision-making model for a power plant inventory shortage scenario is constructed. By calculating the inventory of the downstream power plant and the latest arrival time of the ship, the port ship scheduling and shipping plan optimization, as well as the ship replacement algorithm, are combined and called.

[0009] Preferably, the process of constructing a decision model for a scenario with insufficient coal loading and then combining and calling algorithms is as follows: Obtain the estimated loading volume and loading plan for each loading station, and calculate the deviation of the number of loading trains between the estimated loading volume and the loading plan; for any day, subtract the planned number of loading trains from the estimated number of loading trains at the loading station to obtain the deviation of the number of loading trains at the corresponding loading station. If the daily loading deviation is not greater than the predetermined loading deviation threshold, the deviation will not affect the execution of the supply chain plan and no adjustment will be made. If the loading deviation on any day exceeds the predetermined loading deviation threshold, the loading optimization algorithm will be called to schedule and adjust the loading volume of each loading station at each time. For any day, the number of trains loaded at all loading stations on that day will be added up to obtain the loading number of the corresponding loading station. The loading numbers of all loading stations will be added up to obtain the optimized number of coal trains leaving the area for that day. The planned loading numbers of all loading stations on that day will be added up to obtain the planned number of trains leaving the area. The algorithm compares the number of coal trains leaving the area with the planned number of trains leaving the area. If they are equal, no adjustment is made. If there is a difference, the algorithm for optimizing the return of coal trains with loaded cargo is called. For any route, the number of trains with loaded cargo that are loaded at the loading station on that route is accumulated to obtain the weight of loaded cargo on the corresponding route. For any route, the weight of loaded cargo is compared with the corresponding downstream loaded cargo diversion plan. If the two are equal, no adjustment is made. If there is a discrepancy, the algorithm for optimizing the downstream loaded cargo diversion of railways is called.

[0010] Preferably, the process of constructing a decision model for a scenario of reduced transport capacity in a railway section and combining and calling algorithms is as follows: For each railway route, the transport capacity at all times on that railway route on that day is accumulated to obtain the transport capacity of the corresponding railway route for that day; the transport capacity of the railway route on that day is compared with the planned transport capacity of the corresponding railway route for that day; when the transport capacity of the railway route is not less than the planned transport capacity of the corresponding railway route, no adjustment is made; When the transport capacity of a railway route is less than the planned transport capacity of the corresponding railway route, the coal train load-to-empty optimization algorithm is invoked; the load weight of each route after the algorithm is invoked is calculated; for any route, the load weight is compared with the corresponding downstream load diversion plan; if the two are equal, no adjustment is made; if there is a deviation, the loading optimization algorithm and the downstream railway load diversion optimization algorithm are invoked to adjust the loading plan and the unloading diversion plan.

[0011] Preferably, the process of constructing a scenario model of decreased heavy vehicle handling capacity at key stations and then combining and calling algorithms is as follows: For each railway station, the number of tracks under maintenance at the station is divided by the total number of tracks at the station to obtain the degree of reduction in the heavy-load handling capacity of the station. For each railway route, the correlation matrix between the railway station under maintenance and the route is multiplied by the degree of reduction in the heavy-load handling capacity of the station and the original capacity of the route to obtain the capacity deviation. The original capacity of the railway route is subtracted from the corresponding capacity deviation to obtain the capacity of the railway route at the time of maintenance. When the transport capacity of the railway route during maintenance is less than the operation plan, that is, when there is a deviation between the transport capacity and the train operation plan, the coal train load-to-empty optimization algorithm is called. When the transport capacity of a railway route is less than the planned transport capacity of the corresponding railway route, i.e., there is a deviation between the actual transport capacity and the train operation plan, the coal train load-to-empty-train optimization algorithm is invoked; the load weight of each route after the algorithm is invoked is calculated; for any route, the load weight of the train is compared with the corresponding downstream load diversion plan; if the two are equal, no adjustment is made; if there is a deviation, the loading optimization algorithm and the downstream railway load diversion optimization algorithm are invoked to adjust the loading plan and the unloading diversion plan.

[0012] Preferably, the process of constructing a decision model for abnormal port unloading scenarios and combining and calling algorithms is as follows: The ratio of the number of tipplers under maintenance to the total number of tipplers at the corresponding port is used as the degree of decline in the heavy-load unloading capacity of the port's front station. The port's maintenance schedule is obtained to determine the maintenance status of the port at each time. For any given time, the degree of decline in the heavy-load unloading capacity of the port's front station under maintenance at that time is multiplied by the expected unloading volume of the port's front station at that time, which is used as the unloading volume difference. The expected unloading volume of the port's front station is subtracted from the unloading volume difference to obtain the unloading volume at the corresponding time during maintenance. Compare the unloading volume at the port's forward terminal during maintenance with the original planned port unloading volume; if the unloading volume during maintenance is not less than the original planned port unloading volume, no adjustment will be made. If the unloading volume during maintenance is less than the original port unloading plan, the downstream railway heavy car diversion optimization method is invoked to optimize the port's unloading diversion plan; for each route, the actual number of cross-passage and diversion trains after invoking the downstream railway heavy car diversion plan is calculated. The actual number of trains crossing and diverting on each route is compared with the original number of trains crossing and diverting. If they are the same, no adjustment is made. If they are different, the coal train crossing and diverting optimization algorithm and the loading optimization algorithm are called.

[0013] Preferably, the process of calculating the actual number of overlapping and diverted trains after invoking the downstream railway heavy-car diversion plan is as follows: For each time point, the total number of trains unloading at the port at that time is summed to obtain the planned unloading volume of the corresponding port at that time. For each route, the planned port unloading volume after calling the downstream heavy car diversion plan is subtracted from the original planned port unloading volume, and then multiplied by the correlation coefficient between the port and the corresponding route to obtain the traffic diversion deviation of the corresponding route. The original number of traffic diversion trains on the route is added to the corresponding traffic diversion deviation to obtain the actual number of traffic diversion trains on the corresponding route.

[0014] Preferably, the process of constructing a decision model for a port maritime closure scenario and combining and calling algorithms is as follows: Obtain information on the closure time and determine the closure warning to identify whether it is a maritime closure scenario. If it is a maritime closure scenario, invoke the vessel replacement optimization method and the port vessel scheduling and shipping plan optimization algorithm. Based on the port closure time information, calculate the port inventory at each moment; if the port inventory does not exceed the port safety stock limit, no adjustment is made; if the port inventory exceeds the port safety stock limit, call the downstream railway heavy car diversion optimization algorithm; calculate the actual number of heavy car diversion trains after calling the downstream railway heavy car diversion plan; compare the actual number of heavy car diversion trains on each route with the original number of heavy car diversion trains; if they are the same, no adjustment is made; if they are different, call the coal train heavy car return empty optimization algorithm and loading optimization algorithm.

[0015] Preferably, the process of calculating the port inventory at each moment is as follows: For any given time, multiply the original planned port unloading volume by the average train load capacity to obtain the planned port unloading weight. For each berth, multiply the total number of ships loaded at that berth that did not depart before the start of the closure by the corresponding ship's load capacity as the numerator, and use the transshipment rate of the corresponding berth as the denominator to obtain the loading time of that berth before the end of the closure. Add the loading time to the start time of the closure to obtain the loading end time of the corresponding berth during the closure. For any time before the loading end time or after the end of the closure, use the total loading volume of the berth as the total loading volume of the corresponding berth at that time. For any given time, subtract the total loading volume of all berths at that time from the planned port unloading weight to obtain the port inventory at that time.

[0016] Preferably, the process of constructing a decision model for a power plant inventory shortage scenario and combining and calling algorithms is as follows: For each hydroelectric power plant at any given time, the total deadweight of all ships that arrived at and were associated with the hydroelectric power plant within the time step before the corresponding time is accumulated, which is taken as the unloading volume of the corresponding hydroelectric power plant at the corresponding time. For each hydroelectric power plant, the inventory of the corresponding hydroelectric power plant at the current time is obtained by subtracting the coal consumption within the time step from the inventory of the corresponding hydroelectric power plant at the current time and adding the unloading volume. No adjustment is made if the current inventory of the hydropower plant is not lower than the predetermined safety inventory of the power plant; if it is lower than the predetermined safety inventory of the power plant, the initial inventory is subtracted from the predetermined safety inventory of the power plant and then divided by the coal consumption within the time step to calculate the latest arrival time of the ship to avoid this situation and meet the predetermined safety inventory of the power plant; if the latest arrival time of the ship is greater than the total time of the power plant's route, the port ship scheduling and shipping plan optimization algorithm is called; if the latest arrival time of the ship is not greater than the total time of the power plant's route, the ship replacement algorithm is called.

[0017] A second aspect of the present invention provides an integrated coal supply chain scheduling decision system, using an integrated coal supply chain scheduling decision method, comprising: The basic scheduling optimization algorithm construction module builds a variety of basic scheduling optimization algorithms covering the coal supply chain from railway transportation to port shipping; The coal loading insufficient scenario algorithm calling module combines and calls various basic scheduling optimization algorithms for different scenarios; when the expected loading volume of upstream loading stations is lower than the original loading plan, a coal loading insufficient scenario decision model is constructed; by calculating the number of coal trains leaving the area, the deviation of the number of loading trains at loading stations, and the weight of loaded cars, the coal train loading optimization, the loaded return empty optimization, and the downstream heavy car diversion optimization algorithms are combined and called. The railway section capacity reduction scenario algorithm calling module calculates the railway route's capacity for the day based on the real-time capacity of each railway section; when the day's capacity is less than the planned capacity, it constructs a railway section capacity reduction scenario decision model; by calculating the railway route's capacity and the weight of loaded cars, it combines and calls algorithms for coal train loaded return to empty, loading optimization, and downstream loaded car diversion optimization. The module for calling the algorithm for the scenario of reduced heavy car handling capacity at key stations constructs a scenario model when the unloading capacity of a station is lower than the predetermined unloading capacity. It combines and calls algorithms for optimizing the return of coal trains to empty trains, loading, and downstream heavy car diversion by calculating the transport capacity and the weight of heavy cars delivered during railway maintenance. The port unloading anomaly scenario algorithm calling module constructs a port unloading anomaly scenario decision model when the port's unloading capacity is lower than the predetermined capacity; it combines and calls algorithms for optimizing downstream heavy car diversion, optimizing coal trains returning empty after being loaded, and optimizing loading by calculating the port's unloading volume during maintenance and the actual number of trains diverted for heavy traffic. The port shipping closure scenario algorithm calling module constructs a port shipping closure scenario decision model when the port prohibits ships from entering or leaving; by calculating the port inventory and the actual number of trains diverted and loaded, it combines and calls the downstream railway heavy car diversion optimization algorithm, the coal train heavy car return empty optimization algorithm, and the loading optimization algorithm. The power plant inventory shortage scenario algorithm calling module constructs a power plant inventory shortage scenario decision model when a ship malfunction causes the corresponding power plant to be unable to receive coal on time or the inventory is lower than the predetermined power plant safety inventory. By calculating the inventory of the downstream power plant and the latest arrival time of the ship, it combines and calls the port ship scheduling and shipping plan optimization and ship replacement algorithm.

[0018] A third aspect of the present invention provides a terminal, including a processor and a storage medium; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute steps according to an integrated coal supply chain scheduling decision-making method.

[0019] A fourth aspect of the invention provides a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the steps of an integrated coal supply chain scheduling decision-making method.

[0020] Compared with the prior art, the beneficial effects of the present invention include at least the following: 1. This invention establishes a multi-scenario collaborative scheduling system covering the entire coal supply chain, including upstream loading, midstream railway transportation, downstream port unloading and maritime transport, and power plant inventory. By identifying various disturbance scenarios such as insufficient loading, reduced railway capacity, insufficient handling capacity at key stations, abnormal port unloading, port closures, and power plant inventory shortages, the system automatically combines and applies multiple fundamental algorithms, including loading optimization, load-to-load return optimization, downstream diversion optimization, vessel scheduling optimization, and vessel replacement, to achieve rapid supply chain response.

[0021] 2. This invention achieves a gradual diffusion of scheduling adjustments from local to upstream and downstream processes by detecting loading deviations, transport capacity deviations, unloading capacity deviations, diversion deviations, and inventory deviations at each level. In each scenario, optimization is prioritized at a single local link. Only when local adjustments fail to meet planned requirements are higher-level coordinated optimizations triggered across the entire chain. This effectively avoids the problem of a single change causing large-scale adjustments across the entire chain, significantly reducing the propagation range of scheduling disturbances and improving the overall stability of the plan.

[0022] 3. This invention achieves precise matching and optimized allocation of coal transportation resources (such as railway capacity, station track capacity, port unloading capacity, port inventory, and power plant inventory) by dynamically calculating key indicators such as route capacity, station handling capacity, port unloading capacity, port inventory, and power plant inventory. Through the synergistic effect of multi-scenario optimization algorithms, it can effectively reduce problems such as train backlogs, uneven loading and unloading, and resource waste caused by port congestion, railway bottlenecks, shipping closures, and inventory shortages, thereby improving overall transportation efficiency, ensuring stable supply to power plants and ports, and enhancing the economy and security of the coal supply chain. Attached Figure Description

[0023] Figure 1 A flowchart of an integrated coal supply chain scheduling decision-making method provided by the present invention; Figure 2 This is a flowchart of the coal train load-shifting and empty-train optimization algorithm in an embodiment of the present invention; Figure 3 This is a flowchart of the ship replacement optimization algorithm in an embodiment of the present invention; Figure 4 This is a flowchart of the decision-making model for a scenario with insufficient coal loading in an embodiment of the present invention; Figure 5 This is a flowchart of the decision-making model for a scenario of reduced transport capacity in a railway section, as described in this embodiment of the invention. Figure 6 This is a flowchart of a decision-making model for a scenario where the heavy vehicle handling capacity at a key station decreases, as described in an embodiment of the present invention. Figure 7 This is a flowchart of the decision-making model for abnormal port unloading scenarios in an embodiment of the present invention; Figure 8This is a flowchart of the port maritime closure scenario decision model in an embodiment of the present invention; Figure 9 This is a flowchart of the power plant inventory shortage decision-making model in an embodiment of the present invention; Figure 10 This is a schematic diagram of an integrated coal supply chain in an embodiment of the present invention; Figure 11 This is a line graph showing the port inventory situation before and after scheduling in an embodiment of the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of this invention.

[0025] Example 1 Embodiment 1 of the present invention provides an integrated coal supply chain scheduling decision-making method, see below. Figure 1 This includes the following steps: S1. Construct a variety of basic scheduling optimization algorithms covering the coal supply chain from railway transportation to port shipping, including coal train loading optimization algorithm, coal train load-and-load return optimization algorithm, railway downstream loaded car diversion optimization algorithm, port ship queuing and shipping plan optimization algorithm, and ship replacement optimization algorithm.

[0026] As a preferred implementation method, the coal loading optimization algorithm takes minimizing loading costs and minimizing the change in the number of coal-carrying trains leaving the area as optimization objectives, and constructs an integer programming model considering the loading capacity and response rate constraints of each loading station. Construct an objective function for loading optimization that considers both the coal outbound demand satisfaction rate and loading costs:

[0027] In the formula, The objective function for loading is optimized; i represents the i-th station, j represents the j-th train, t represents the time, and t0 is the initial time. This represents the load capacity of the j-th train; The variable is 0 / 1, representing whether the j-th train is loaded at time t in the i-th station, where 0 indicates no loading and 1 indicates loading. This indicates the total demand for loading and unloading within the region during the period. This represents the total number of vehicles loaded and shipped out of the area within time t; This is an adjustment factor used to measure the importance of meeting coal demand in the region and loading costs. βThis is an adjustment coefficient used to adjust the importance of the total outbound volume and the outbound volume satisfaction rate per unit time within the optimization cycle; The loading cost, which indicates whether train j is loaded at station t at time i, is related to the distance between the coal mine and the station and the loading operation method. Loading constraints include the loading capacity constraints and response rate constraints of the stations; specifically, they are as follows: ; In the formula, the first term is the loading capacity constraint, which is affected by the station's supporting facilities; the second term is the response rate capacity constraint, which is affected by the dispatch rate of station personnel or equipment. This represents the loading capacity of the i-th station at time t; Let represent the responsive loading rate at time t for the i-th station.

[0028] As a preferred implementation method, the coal train load-to-empty-train optimization algorithm is mainly an optimization scheduling algorithm for allocating loaded cars to various railway lines and empty cars to various loading areas. The load-transfer algorithm and the empty-train return algorithm are logically similar, allocating load-transfer and empty-train return plans according to the diversion plan or loading plan. The following uses the empty-train return algorithm as an example... Figure 2 The steps are illustrated with specific formulas: Based on the loading capacity of each station, the correlation matrix between the station and the empty return transport route, and the number of empty return trains on each empty return transport route, it is determined whether the empty return plan for each region meets the loading plan requirements; the specific formula is as follows: ; In the formula, This represents the value in the i-th row and l-th column of the association matrix between the loading station and the return empty transport route, i.e., the value of the loading station. i and return air transport route l The correlation between them; if the return path l passes through loading station i, it is 1; if it does not pass through, it is 0. express t Time l The number of empty trains returning to their original routes; If the return empty plan for each region meets the loading plan requirements, determine whether the total return empty plan meets the total loading plan requirements; the specific formula is as follows: ; If the total empty return plan meets the total loading plan requirements, the empty return plan will be allocated proportionally, as shown in the following formula: ; If the total empty return plan does not meet the total loading plan demand, a quadratic programming model is constructed under the constraint of transport capacity. The objective is to achieve even loading distribution across all routes, constrained by railway capacity limitations, to adjust the empty return plan. The objective function is expressed as: ; The constraints are expressed as follows: ; In the formula, The objective function is... Indicates the first l The maximum railway capacity of a railway route at any given time.

[0029] The downstream heavy-car diversion optimization algorithm for railways primarily focuses on the formulation of diversion plans for downstream sections of coal railway transportation. Currently, the decline in railway transportation efficiency is often due to heavy-car congestion before ports. Therefore, rationally scheduling heavy-car operations to prevent port congestion is a major concern for railway dispatchers. Port congestion is often caused by unreasonable downstream heavy-car diversion plans, resulting in excessively high train density at the port during a certain period, exceeding the port's unloading capacity. The algorithm aims to minimize the retention time of heavy-cars before the port or along the railway line. The retention time of trains in railway transportation directly reflects the railway's transport efficiency. Constraints include plan balance constraints, port inventory and unloading capacity constraints, railway transport capacity and station transport capacity constraints, and heavy-car operation constraints.

[0030] As a preferred implementation method, the downstream heavy vehicle diversion optimization algorithm for railways aims to maximize the turnover efficiency of heavy vehicles by minimizing the total number of heavy vehicles retained within the optimization period, and constructs a diversion optimization objective function as follows: ; In the formula, Optimize the objective function for stream splitting; i Indicates the first i There are 1 station, with a total of n One station; k Indicates the first k A loaded vehicle, where t represents time; i =1 indicates the originating station. i =n represents the port station. t = t 0 represents the initial time, T represents the termination time, and K0 represents the total number of multiple trains received at the originating station within the optimization period; express t Time of the first j Are the heavy vehicles in i The station will be retained. The constraints of the downstream heavy car diversion optimization algorithm for railways include planning balance constraints, port inventory and unloading capacity constraints, railway transport capacity and station transport capacity constraints, and heavy car operation constraints. The planned balance constraint states that the sum of the departure schedule from the originating station and the number of loaded cars en route at the beginning of the decision period must equal the sum of the number of cars unloaded at each station during the decision period and the number of trains remaining on the line at the end of the period; expressed as follows: ; In the formula, Indicate whether the k-th loaded vehicle passes through station i at time t; Indicates whether the k-th loaded car is diverted or unloaded at station i at time t; This represents the number of loaded trains en route at the i-th station at the start of the optimization. Port inventory and unloading capacity constraints are: ; In the formula, This represents the port inventory at time t; For time step; This represents the amount of coal transported by the kth loaded vehicle; This represents the port loading volume at time t; , This indicates the upper and lower limits of the port's safety stock.

[0031] The constraints on railway transport capacity and station transport capacity are as follows: ; In the formula, This represents the maximum number of loaded cars allowed to pass through the i-th station during time period t; This represents the number of loaded cars that can be reserved at the i-th station; The loaded vehicle operation constraint indicates that each car can only be unloaded at one station. At any station and at any time, a loaded car can only be in one of the following states: held, running, or unloading, or not in any of these states. The corresponding constraint is: .

[0032] As a preferred implementation method, the port vessel scheduling and shipping plan optimization algorithm mainly targets scenarios such as port destocking and power plant supply assurance, optimizing the vessels and timing of vessel scheduling adjustments. Considering the port's vessel turnover efficiency and minimizing vessel plan adjustments, a vessel scheduling optimization objective function is constructed. The objective of maximizing the loading volume within the optimization period reflects the port's turnover efficiency, while the objective of minimizing the change in the loading sequence of port vessels within the optimization period reflects the plan adjustment amount. The vessel scheduling optimization objective function is expressed as: ; In the formula, To optimize the objective function for arranging boats, , , The weighting coefficient is determined based on experience and parameter tuning. In this embodiment, the weights are generally set to 0.8, 0.1, and 0.1, depending on the importance placed on the task by the dispatcher. This represents the target for optimizing port turnover efficiency of ships, and is used to calculate the amount of cargo loaded within the statistical period. Let be a continuous variable, representing the loading capacity of the a-th ship; This indicates the original plan for ships that have already been assigned a berth order, i.e., ship a was originally scheduled to be loaded at berth b in order c. This indicates the cost of adjusting the original order of the ships; This refers to the cost of transferring ships that are already in the queue to the next waiting queue. The variable is 0 / 1, indicating whether the a-th ship is loaded at the b-th berth in the c-th order; The constraints of the port vessel scheduling and shipping plan optimization algorithm include emergency power plant safety stock constraints, vessel type and berth matching constraints, vessel arrival and departure time constraints, and vessel sequential berthing constraints, which are specifically expressed as follows: ; ; ; ; In the formula, The variable is 0 / 1, indicating whether the a-th ship is loaded at the b-th berth in the c-th order and dispatched before the closure of the port. 1 indicates yes and 0 indicates no. The variable is 0 or 1, indicating whether the power plant corresponding to the a-th ship is in emergency, with 1 indicating yes and 0 indicating no. The allowable load capacity for the b-th berth. The deadweight of the a-th vessel; This is a maximum value, used as an auxiliary constant for linearizing the conditional constraints; when At that time, it indicates that ship A is docked at berth B. This means the ship's deadweight is lower than the berth's allowable deadweight. When, it means that vessel A is not berthed at berth B. This means that if the difference between the ship's deadweight and the berth's allowable deadweight is less than a maximum value, no relevant constraints will be imposed. This indicates the loading time of the a-th vessel at the b-th berth; This indicates the departure time of the ship currently being loaded at berth b; This indicates the start time when vessels are prohibited from entering or leaving the port.

[0033] As a preferred implementation method, the vessel replacement optimization algorithm is mainly designed to address scenarios where power plants are facing severe inventory shortages. This requires adjusting the destinations of vessels en route to resolve the shortage, a scenario where adjustments to shipping schedules are not feasible. The algorithm replaces vessels en route based on the region of the power plant in distress, its affiliated subsidiary, and the inventory status of other power plants, adhering to the principle of minimizing adjustment costs. In this embodiment, since the communication costs of adjustments cannot be quantified, priority is given to replacing power plants belonging to the same company. Figure 3 As shown, the algorithm's logical steps are as follows: Based on the loading capacity of each ship, the initial inventory of the power plant, and the coal consumption rate, calculate the power plant's inventory at different times; the formula is shown below: ; In the formula, This represents the inventory of hydropower plant p at time t; For time step; This represents the coal consumption of the hydropower plant p within a time step, i.e., the coal consumption rate. Let p be the value in the a-th row and p-th column of the correlation matrix between the hydropower plant and the ship, representing the correlation degree between the hydropower plant p and the ship a. A value of 1 indicates the ship passes through the hydropower plant, and a value of 0 indicates the ship does not pass through the hydropower plant. Indicates whether ship a t - The arrival time of the ship at the hydroelectric power plant (i.e., the destination) after time t and before time t is a 0 / 1 variable, where 0 represents no and 1 represents yes. The ship's arrival time is calculated by adding the route running time and the ship's departure time. If the power plant's inventory is lower than the predetermined safety stock, calculate the latest arrival time when the power plant reaches the predetermined safety stock; the calculation formula is as follows: ; In the formula, The latest ship arrival time for the power plant p, which is in a state of emergency, is required to meet the planned safety stock needs of the power plant. This represents the initial inventory of the hydroelectric power plant p; The planned safety stock for the downstream hydroelectric power plant p; Calculate the difference between the ship's route travel time and the current travel time as the arrival time of the corresponding ship en route to the power plant; iterate through all ships en route and determine whether the arrival time to the power plant is less than the arrival time of the latest required ship; when there is a ship whose arrival time to the power plant is less than the arrival time of the latest required ship, determine whether the corresponding ship and the power plant to which the replacement ship belong to belong to the same secondary company according to time priority (i.e., arrival time to the power plant from earliest to latest); if so, then perform ship replacement.

[0034] S2. Based on typical scenarios in the scheduling process, the basic optimization algorithms are combined and called to construct decision optimization models under typical scenarios. The combined calling logic algorithm is shown in the calling process of each scenario below.

[0035] Based on typical scenarios in the scheduling process, basic optimization algorithms are combined and invoked to construct decision optimization models for typical scenarios. Referring to Table 1, typical scenarios include insufficient coal loading, reduced railway capacity, decreased capacity for receiving and dispatching loaded wagons at key stations, abnormal port unloading, port closures, power plant inventory shortages due to ship malfunctions, and abnormal power plant inventory due to changes in coal consumption. Generally, it is rare for multiple scenarios to occur simultaneously. When multiple scenarios occur simultaneously in extreme cases, the scheduling principle of prioritizing supply and transportation and minimizing the impact on coal production is followed. The models are invoked sequentially from downstream to upstream, i.e., gradually applying models for scenarios such as power plant inventory shortages, port closures and unloading abnormalities, reduced railway capacity, and insufficient coal loading.

[0036] Table 1. Typical Scenarios

[0037] As a preferred implementation, when the expected loading volume at an upstream loading station is lower than the original loading plan, it indicates insufficient coal loading. A decision-making model for insufficient coal loading scenarios is adopted. Based on the principle of minimizing the impact, scheduling adjustments are gradually diffused downstream. The decision-making model for insufficient coal loading scenarios is designed, and the algorithm combination and invocation strategy is as follows: Figure 4 As shown; by calculating the number of coal trains leaving the area, the deviation in the number of loading trains at loading stations, and the weight of loaded cars, the algorithms for optimizing coal train loading, optimizing empty return trips, and optimizing the diversion of loaded cars downstream of the railway are combined and called; the specific calling process is as follows: Obtain the estimated loading volume and loading plan for each loading station, and calculate the deviation in the number of loading trains between the estimated loading volume and the loading plan. For any given day, subtract the planned number of loading trains from the estimated number of loading trains at the loading station to obtain the corresponding deviation in the number of loading trains at the loading station. The specific formula is as follows: ; In the formula, d represents the d-th day; This indicates the deviation in the number of trains loaded at the i-th loading station on day d; This represents the number of loading schedule columns for the i-th loading station on day d; This represents the number of trains expected to be loaded at the i-th loading station on day d. If the number of daily loading deviations does not exceed the predetermined loading deviation threshold, the deviation will not affect the execution of the supply chain plan and no adjustments will be made. If the deviation in the number of loading trains on any given day exceeds the predetermined loading deviation threshold, the loading optimization algorithm is invoked to adjust the loading volume at each loading station at each time. For any given day, the number of trains loading at all loading stations on that day is summed to obtain the number of loading trains at that station. The number of loading trains at all loading stations is summed to obtain the optimized number of coal trains leaving the area for that day. The planned number of loading trains at all loading stations on that day is summed to obtain the planned number of trains leaving the area. The specific formula is as follows: ; In the formula, t represents the t-th moment of day d; This represents the number of loading trains after optimization at the i-th loading station on day d; To determine whether train j is loaded at station ith at time t; To optimize the number of coal trains leaving the area on day d; This represents the planned outbound column number for day d. This represents the planned number of trains to be loaded at the i-th loading station on day d. Compare the number of coal trains leaving the area with the planned number of trains leaving the area; if they are equal, no adjustment is made. If discrepancies exist, the coal train load-and-load return optimization algorithm is invoked. For any given path, the number of loaded cars loaded at the loading station on that path that undergo load-and-load transfers is accumulated to obtain the load-and-load transfer weight for that path, which is the total downstream diversion weight for each path. The calculation method for the number of load-and-load transfer trains on each path is shown in the following formula, that is, the number of load-and-load transfer trains on each path is equal to the number of times all trains at all loading stations run on the corresponding path on that day: ; In the formula, This represents the weight of loaded vehicles on route l on day d, which is the total weight of loaded vehicles diverted downstream of route l. The variable is 0 / 1, indicating whether the j-th train crosses the path l. For any path, compare the weight of loaded vehicles with the corresponding downstream loaded vehicle diversion plan; if the two are equal, no adjustment is made; if there is a deviation, call the railway downstream loaded vehicle diversion optimization algorithm.

[0038] As a preferred implementation, the number of loaded trains on each route can also be obtained by summing the number of loaded trains unloaded at all stations on the same day and the number of trains that are loaded on the corresponding route; it is expressed as follows, that is, the number of loaded trains on each route is equal to the total number of trains that unload or divert at each unloading station and pass through the corresponding route: ; In the formula, The variable is 0 / 1, indicating whether the k-th train crosses the l-th path; Indicates whether the k-th train is unloaded at time t at the i-th station; Because trains are grouped at the upstream and downstream junctions during railway transportation, upstream and downstream trains are represented by different symbols j and k when constructing the model, but the total number of trains is balanced.

[0039] As a preferred implementation method, the daily transport capacity of the railway line is calculated based on the real-time transport capacity of each railway segment. When the daily transport capacity is less than the planned transport capacity, it indicates a decrease in the railway line's transport capacity. A decision-making model for a railway segment's transport capacity decrease scenario is adopted. Based on the principle of minimizing the impact, scheduling adjustments are gradually diffused downstream. The algorithm combination and invocation strategy for the decision-making model for a railway segment's transport capacity decrease scenario is designed as follows: Figure 5 As shown; by calculating the railway route's transport capacity and the weight of loaded cars, the algorithms for optimizing coal trains' return to empty tracks with loaded cars, loading optimization, and downstream loaded car diversion optimization are combined and called; the specific calling process is as follows: For each railway route, the transport capacity at all times on that route on that day is summed up to obtain the transport capacity for that railway route on that day; the calculation formula is as follows: ; In the formula, This represents the transport capacity on the l-th railway route on day d; Let represent the transport capacity of the l-th railway transport route at time t. Due to factors such as weather, the railway's transport capacity is affected. It will change; The capacity of the railway route on that day is compared with the planned capacity of the corresponding railway route on that day; no adjustment is made when the capacity of the railway route is not less than the planned capacity of the corresponding railway route. When the transport capacity of a railway route is less than the planned transport capacity of the corresponding railway route, i.e., there is a deviation between the actual transport capacity and the train operation plan, the coal train load-and-return optimization algorithm is invoked. For any route, the number of trains carrying loaded cars that are loaded at the loading station on that route is accumulated to obtain the load-and-return weight of the corresponding route. For any route, the load-and-return weight is compared with the corresponding downstream load-and-return plan. If the two are equal, no adjustment is made. If there is a deviation, the loading optimization algorithm and the downstream load-and-return optimization algorithm are invoked to further adjust the loading plan and the unloading and diversion plan.

[0040] As a preferred implementation, when the unloading capacity of a station is lower than the predetermined unloading capacity, it indicates a decrease in the loaded train handling capacity of key stations. A scenario model of decreased loaded train handling capacity at key stations is adopted. Supply chain simulation or actual operation may experience segmental congestion. Based on the principle of minimizing the impact, scheduling adjustments are gradually diffused upstream and downstream. The algorithm combination and invocation strategy for this model is designed as follows: Figure 6As shown; by calculating the railway route's capacity and the weight of loaded cars during railway maintenance, the algorithms for optimizing coal trains' return to empty tracks with loaded cars, loading optimization, and downstream loaded car diversion optimization are combined and called; the specific calling process is as follows: For each railway station, the number of tracks under maintenance at the station is divided by the total number of tracks at that station to obtain the degree of reduction in the station's loaded train handling capacity. For each railway route, the correlation matrix between the railway station under maintenance and the route is multiplied by the degree of reduction in the station's loaded train handling capacity and the route's original capacity to obtain the capacity deviation. The original capacity of the railway route is subtracted from the corresponding capacity deviation to obtain the railway route's capacity at the time of maintenance. The specific formula is shown below: ; In the formula, This represents the number of tracks at the i-th railway station; Indicates the number of tracks under maintenance; This indicates the percentage of the railway station's capacity to handle loaded trains that has decreased. This represents the original transport volume of railway route l at time t; This indicates the transport capacity of railway route l after the change caused by the decrease in the unloading capacity of stations along route l; Let i be the value in the i-th row and l-th column of the association matrix between railway station i and path l, i.e., the value of the railway station. i and railway routes l The correlation between them; if path l passes through railway station i, it is 1; if it does not pass through, it is 0. The variable is 0 or 1, indicating whether the i-th station is under maintenance at time t, where 1 represents yes and 0 represents no. When the transport capacity of the railway route during maintenance is less than the operation plan, that is, when there is a deviation between the transport capacity and the train operation plan, the coal train load-to-empty optimization algorithm is called. When the transport capacity of a railway route is less than the planned transport capacity of the corresponding railway route, i.e., there is a deviation between the actual transport capacity and the train operation plan, the coal train load-and-return optimization algorithm is invoked. For any route, the number of trains carrying loaded cars that are loaded at the loading station on that route is accumulated to obtain the load-and-return weight of the corresponding route. For any route, the load-and-return weight is compared with the corresponding downstream load-and-return plan. If the two are equal, no adjustment is made. If there is a deviation, the loading optimization algorithm and the downstream load-and-return optimization algorithm are invoked to further adjust the loading plan and the unloading and diversion plan.

[0041] As a preferred implementation, when the port's unloading capacity is lower than the predetermined capacity, it indicates a port unloading anomaly. A port unloading anomaly scenario decision model is adopted. Since the unloading anomaly mainly affects railway unloading, based on the principle of minimizing the impact range, scheduling adjustments are gradually propagated upstream. The algorithm combination and invocation strategy of this model is designed as follows: Figure 7 As shown; by calculating the unloading volume at the port during maintenance and the actual number of trains diverted for heavy traffic, the algorithms for optimizing downstream heavy car diversion, optimizing coal trains returning empty after heavy traffic, and optimizing loading are combined and called; the specific calling process is as follows: For each port, the ratio of the number of tipplers under maintenance to the total number of tipplers at the corresponding port is used as the degree of decline in the heavy-load unloading capacity of the port's front-end station. The port's maintenance schedule is obtained to determine the maintenance status at each time point. For any given time, the degree of decline in the heavy-load unloading capacity of the port's front-end station under maintenance at that time is multiplied by the expected unloading volume of the port's front-end station at that time, resulting in the unloading volume difference. The expected unloading volume of the port's front-end station is subtracted from the unloading volume difference to obtain the unloading volume during maintenance at that time. The specific formula is shown below: ; In the formula, This indicates the number of tipplers at a port. This indicates the number of tipplers under maintenance; This indicates the percentage of the port's front-end station that has experienced a decline in its capacity to handle heavy vehicles. This represents the expected unloading volume at the port's forward terminal at time t. This indicates the port's unloading volume after the change caused by the decrease in the port's unloading capacity. The variable is 0 or 1, indicating whether the port is under maintenance at time t, where 1 represents yes and 0 represents no. Compare the unloading volume at the port's forward terminal during maintenance with the original planned port unloading volume; if the unloading volume during maintenance is not less than the original planned port unloading volume, no adjustment will be made. If the unloading volume during maintenance is less than the original port unloading plan, the downstream heavy car diversion optimization method of the railway will be used to optimize the port's unloading diversion plan under the new port unloading capacity constraints. For each time moment, the total number of trains unloading at the port at that time is summed to obtain the planned unloading volume for that port at that time; the specific calculation method is as follows: ; In the formula, This represents the planned unloading volume at the port after invoking the downstream heavy car diversion optimization method of the railway. It is a 0 / 1 variable, indicating whether the k-th train is unloaded at the port at time t; For each route, the planned port unloading volume after invoking the downstream railway heavy car diversion plan is subtracted from the original planned port unloading volume, and then multiplied by the correlation coefficient between the port and the corresponding route to obtain the traffic diversion deviation for the corresponding route. The original number of traffic diversion trains on the route is added to the corresponding traffic diversion deviation to obtain the actual number of traffic diversion trains on the corresponding route after invoking the downstream railway heavy car diversion plan. The calculation method is as follows: ; In the formula, This represents the actual number of trains crossing or diverting on the l-th path at time t; This represents the number of originally scheduled trains crossing or diverting on the l-th path at time t; The original planned port unloading volume at time t represents the unloading volume originally intended for the railway at the port. The value in the l-th column of the port-path association matrix represents the association coefficient between the port and the l-th path. A value of 1 indicates that the port is on the l-th path, and a value of 0 indicates that it is not. The actual number of trains crossing and diverting on each route is compared with the original number of trains crossing and diverting. If they are the same, no adjustment is made. If they are different, it means that the downstream diversion plan has deviated, which will affect the upstream loading plan. The coal train crossing and returning empty optimization algorithm and the loading optimization algorithm are then called.

[0042] As a preferred implementation, when a port prohibits ships from entering or leaving, it signifies a port closure for maritime transport. A port closure scenario decision model is employed to address the abnormal inventory levels at port yards and power plants caused by the closure. Based on the principle of minimizing the impact, scheduling adjustments are gradually propagated downstream. The algorithm combination and invocation strategy for this model is designed as follows: Figure 8 As shown; by calculating port inventory and the actual number of trains diverted for heavy traffic, the optimization algorithms for downstream heavy car diversion, coal train heavy-car return-to-empty-train conversion, and loading optimization are combined and called; the specific calling process is as follows: Obtain information on the closure time based on the closure notice issued by the Maritime Safety Administration, and determine the closure warning to determine whether it is a maritime closure scenario. If it is a maritime closure scenario, call the vessel replacement optimization method and the port vessel scheduling and shipping plan optimization algorithm to readjust the vessel scheduling plan and port berthing plan to adapt to the closure situation. For any given time, multiply the original planned port unloading volume by the average train load capacity to obtain the planned port unloading weight. For each berth, multiply the total number of ships loaded at that berth that did not depart before the start of the closure by the corresponding ship's load capacity as the numerator, and use the transshipment rate of the corresponding berth as the denominator to obtain the loading time of that berth before the end of the closure. Add the closure start time to the loading time to obtain the loading end time of the corresponding berth during the closure. For any time before or after the loading end time, use the total loading volume of the berth as the total loading volume of the corresponding berth at that time. For any given time, subtract the total loading volume of all berths at that time from the planned port unloading weight to obtain the port inventory at that time. The specific formula is shown below: ; In the formula, Let be the port inventory at time t; This represents the total loading volume at the b-th berth; This represents the average train load capacity. Since the transport volume of railway transportation plans in the supply chain is often measured in units of trains, the average train load capacity is often used to convert the tonnage plan for each diversion. This indicates whether berth b is idle at time t, and is a 0 / 1 variable, where 1 represents yes and 0 represents no. , Indicates the start and end times of the closure; The loading time at berth b before the end of the closure period; The variable is 0 / 1, indicating whether the a-th ship is loaded at the b-th berth in the c-th order; The variable is 0 / 1, indicating whether the a-th vessel is loaded at the b-th berth in the c-th order and dispatched before the closure begins; Let be the deadweight of the a-th ship; This indicates the loading rate at berth b; No adjustments will be made when the port inventory does not exceed the port safety stock limit; When the port inventory exceeds the port safety stock limit, it is considered that the inventory is too high, and the downstream railway heavy car diversion optimization algorithm is invoked. For each route, the port unloading plan after invoking the downstream railway heavy car diversion plan is subtracted from the original port unloading plan, and then multiplied by the correlation coefficient between the port and the corresponding route, which is used as the traffic diversion deviation for the corresponding route. The original number of traffic diversion trains on the route is added to the corresponding traffic diversion deviation to obtain the actual number of traffic diversion trains on the corresponding route after invoking the downstream railway heavy car diversion plan. The actual number of trains crossing and diverting on each route is compared with the original number of trains crossing and diverting. If they are the same, no adjustment is made. If they are different, it means that the downstream diversion plan has deviated, which will affect the upstream loading plan. The coal train crossing and returning empty optimization algorithm and the loading optimization algorithm are then called.

[0043] As a preferred implementation, when a ship malfunction causes the corresponding power plant to be unable to receive coal on time or a surge in coal consumption leads to inventory falling below the predetermined safety stock level, it indicates that the power plant's inventory is in crisis. A decision-making model for power plant inventory crisis scenarios caused by ship malfunctions or changes in coal consumption is adopted. This primarily affects the corresponding downstream power plant and generally does not impact the upstream and downstream of the supply chain. The algorithm's calling logic is as follows: Figure 9 As shown; by calculating the inventory of the hydropower plant and the latest arrival time of the ship, the port ship scheduling and shipping plan optimization, and the ship replacement algorithm are combined and called; the specific calling process is as follows: For each hydroelectric power plant at any given time, the total deadweight of all vessels that arrived at and are associated with that hydroelectric power plant within the time step preceding that time is accumulated, and this sum is taken as the unloading volume of that hydroelectric power plant at that time. For each hydroelectric power plant, the inventory at the current time is calculated by subtracting the coal consumption within the time step from the inventory at the current time, and then adding the unloading volume. The specific formula is as follows: ; In the formula, This represents the inventory of hydropower plant p at time t; For time step; This represents the coal consumption of the hydropower plant p within a time step, i.e., the coal consumption rate. Let p be the value in the a-th row and p-th column of the correlation matrix between the hydropower plant and the ship. It represents the correlation degree between the hydropower plant p and the ship a. A value of 1 indicates the ship passes by the hydropower plant, and a value of 0 indicates the ship does not pass by the hydropower plant. Let be the deadweight of the a-th ship; Indicates whether ship A is in t - The arrival time of the ship at the hydroelectric power plant (i.e., the destination) after time t and before time t is a 0 / 1 variable, where 0 represents no and 1 represents yes. The ship's arrival time is calculated by adding the route running time and the ship's departure time. Compare the current inventory of the hydropower plant with the predetermined safety inventory of the power plant; if the current inventory is not lower than the predetermined safety inventory of the power plant, no adjustment is made; When the current inventory of the corresponding power plant is lower than the predetermined safety stock of the power plant, the initial inventory minus the predetermined safety stock is divided by the coal consumption within the time step to calculate the latest arrival time to avoid this situation and meet the predetermined safety stock of the power plant. The calculation method can be expressed as: ; In the formula, The hydroelectric power plant, indicating an emergency, needs to meet its safety stock requirements, and the latest ship arrival time is... For the safety stock of the hydroelectric power plant p; This indicates the initial inventory at the hydroelectric power plant; Compare the latest arrival time with the total travel time of the power plant route; when the latest arrival time... If the arrival time is greater than the total travel time of the power plant's route, the port vessel scheduling and shipping plan optimization algorithm is invoked to adjust the port scheduling plan of the corresponding vessel for the power plant so that it can be loaded and dispatched as soon as possible; if the latest arrival time is not greater than the total travel time of the power plant's route, the vessel replacement algorithm is invoked to solve the inventory shortage problem by replacing the vessels of other power plants en route.

[0044] This embodiment addresses different scheduling scenarios, calling different decision optimization models for each scenario to provide auxiliary support for plan adjustments in different scenarios.

[0045] Example 2 Embodiment 2 of the present invention applies an integrated coal supply chain scheduling decision-making method to the multi-stage collaborative transportation business of a large energy company's coal supply chain. To reduce computational scale and for anonymization purposes, the integrated coal supply chain of the energy group is simplified. A simplified schematic diagram of the integrated coal supply chain is shown below. Figure 10 As shown, stations A through N are several railway stations, with station A being the main loading station, stations M and F being the main junctions, and station N being the port's foreground station. Coal is mainly loaded at station A and transported by railway line to users along the line, social ports, and the company's own ports. From the company's own ports, it is loaded onto ships at several berths and shipped to downstream users.

[0046] The following example illustrates a scenario where a maritime closure occurs. Due to weather conditions, the port is expected to close for approximately two days, prohibiting ships from entering or leaving. This will not affect ships already in port that are loading cargo. Since a maritime closure can lead to a shortage of inventory at downstream power plants, a continuous increase in port inventory during the closure period, and a decrease in port unloading capacity, potentially causing a series of problems such as railway congestion before downstream heavy vehicles arrive at the port, the closure and original plan information are input into a maritime closure scenario decision-making model. The model automatically combines algorithms based on this information.

[0047] The port has three berths, all with a capacity of 100,000 tons or more, and its inventory is at a high level, sufficient to meet the loading needs of ships. Currently, each of the three berths has one ship loading: Ship X, Ship Y, and Ship Z. Information on ships en route, in the queue, and waiting in the queue is shown in Table 2 below.

[0048] Table 2 Ship Information Sheet

[0049] By invoking the aforementioned port closure scenario decision model, intelligent optimization decisions for the scenario are achieved. The ship loading sequence and berth allocation before and after scheduling are shown in Table 3. The rescheduled loading sequence and berth allocation fully utilize the gaps during the closure period, ensuring that large ships arrive at the port before the closure and load during the closure period, thus improving port efficiency and minimizing the impact of the closure on loading. Furthermore, the power plants corresponding to ships M and N were swapped. Analysis reveals that the closure of power plant 8 prevented ship M from arriving normally, leading to a shortage of coal. Swapping ship N for power plant 9 ensured a stable coal supply.

[0050] Table 3. Vessel berthing plan at each berth before and after the end of the navigation blockade.

[0051] Table 4. Downstream Heavy Vehicle Diversion Plan Before Scheduling

[0052] Table 5. Downstream Heavy Vehicle Diversion Plan After Scheduling

[0053] The downstream heavy-load train operation diversion plans for the downstream port section of the railway before and after the adjustment are shown in Tables 4 and 5. It can be seen that the adjustment of the train operation plan reduced the number of trains arriving at the port and adjusted the density distribution of the diverted train flow. The port inventory situation during this period is as follows: Figure 11 As shown, this also prevented port inventories from exceeding their maximum warning line.

[0054] Since the closure of the waterway did not affect the upstream vehicle arrival schedule, the inventory levels at the port and power plant returned to normal after optimization, and the model did not call the loading optimization algorithm to adjust the loading. The combination of the above algorithms achieved decision optimization in a real-world scheduling scenario, assisting in the formulation of a joint scheduling plan.

[0055] Example 3 Embodiment 3 of the present invention provides an integrated coal supply chain scheduling decision system, comprising: The basic scheduling optimization algorithm construction module builds a variety of basic scheduling optimization algorithms covering the coal supply chain from railway transportation to port shipping; The coal loading insufficient scenario algorithm calling module combines and calls various basic scheduling optimization algorithms for different scenarios; when the expected loading volume of upstream loading stations is lower than the original loading plan, a coal loading insufficient scenario decision model is constructed; by calculating the number of coal trains leaving the area, the deviation of the number of loading trains at loading stations, and the weight of loaded cars, the coal train loading optimization, the loaded return empty optimization, and the downstream heavy car diversion optimization algorithms are combined and called. The railway section capacity reduction scenario algorithm calling module calculates the railway route's capacity for the day based on the real-time capacity of each railway section; when the day's capacity is less than the planned capacity, it constructs a railway section capacity reduction scenario decision model; by calculating the railway route's capacity and the weight of loaded cars, it combines and calls algorithms for coal train loaded return to empty, loading optimization, and downstream loaded car diversion optimization. The module for calling the algorithm for the scenario of reduced heavy car handling capacity at key stations constructs a scenario model when the unloading capacity of a station is lower than the predetermined unloading capacity. It combines and calls algorithms for optimizing the return of coal trains to empty trains, loading, and downstream heavy car diversion by calculating the transport capacity and the weight of heavy cars delivered during railway maintenance. The port unloading anomaly scenario algorithm calling module constructs a port unloading anomaly scenario decision model when the port's unloading capacity is lower than the predetermined capacity; it combines and calls algorithms for optimizing downstream heavy car diversion, optimizing coal trains returning empty after being loaded, and optimizing loading by calculating the port's unloading volume during maintenance and the actual number of trains diverted for heavy traffic. The port shipping closure scenario algorithm calling module constructs a port shipping closure scenario decision model when the port prohibits ships from entering or leaving; by calculating the port inventory and the actual number of trains diverted and loaded, it combines and calls the downstream railway heavy car diversion optimization algorithm, the coal train heavy car return empty optimization algorithm, and the loading optimization algorithm. The power plant inventory shortage scenario algorithm calling module constructs a power plant inventory shortage scenario decision model when a ship malfunction causes the corresponding power plant to be unable to receive coal on time or the inventory is lower than the predetermined power plant safety inventory. By calculating the inventory of the downstream power plant and the latest arrival time of the ship, it combines and calls the port ship scheduling and shipping plan optimization and ship replacement algorithm.

[0056] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0057] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination of the foregoing. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0058] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0059] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0060] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. An integrated coal supply chain scheduling decision method, characterized in that, Comprise: Construct a variety of basic scheduling optimization algorithms covering coal supply chain railway transportation to port shipping; For different scenarios, call a variety of basic scheduling optimization algorithms in combination; when the expected loading capacity of the upstream loading station is lower than the original loading plan, construct a coal loading shortage scenario decision model; by calculating the coal train loading number, the loading train number deviation of the loading station and the heavy train interchange weight, the coal train loading optimization, heavy train interchange and empty optimization and railway downstream heavy train shunting optimization algorithms are combined and called; Based on the real-time transport capacity of each section of railway line, calculate the transport capacity of the railway line on the same day; When the transport capacity of the same day is less than the planned transport capacity, construct a railway section transport capacity decline scenario decision model; by calculating the transport capacity of the railway path and the heavy train interchange weight, the coal train interchange and empty optimization, loading optimization and railway downstream heavy train shunting optimization algorithms are combined and called; When the unloading capacity of the station is lower than the predetermined unloading capacity, construct a key station heavy train unloading capacity decline scenario model; by calculating the transport capacity of the railway path during railway maintenance and the heavy train interchange weight, the coal train interchange and empty optimization, loading optimization and railway downstream heavy train shunting optimization algorithms are combined and called; When the unloading capacity of the port is lower than the predetermined capacity, construct a port unloading abnormal scenario decision model; by calculating the unloading capacity of the port during maintenance and the actual heavy train interchange and shunting train number, the railway downstream heavy train shunting optimization, coal train interchange and empty optimization and loading optimization algorithms are combined and called; When the port prohibits ships from entering and leaving, construct a port shipping blockade scenario decision model; by calculating the port inventory and the actual heavy train interchange and shunting train number, the railway downstream heavy train shunting optimization, coal train interchange and empty optimization and loading optimization algorithms are combined and called; When there is a ship failure causing the corresponding power plant to be unable to arrive on time or the inventory is lower than the predetermined power plant safety inventory, construct a power plant inventory emergency scenario decision model; By calculating the inventory of the downstream power plant and the longest arrival time, the port ship scheduling and shipping plan optimization and ship replacement algorithm are combined and called.

2. The integrated coal supply chain scheduling decision method according to claim 1, characterized in that: The process of constructing a coal loading shortage scenario decision model for algorithm combination calling is: Obtain the expected loading capacity and loading plan of each loading station, and calculate the loading train number deviation between the expected loading capacity and the loading plan; for any day, subtract the expected achievable loading train number of the loading station from the loading plan train number to obtain the loading train number deviation of the corresponding loading station; If the daily loading deviation train number is not greater than the predetermined loading deviation threshold, the deviation does not affect the supply chain plan execution and no adjustment is made; if there is any day with a loading train number deviation greater than the predetermined loading deviation threshold, call the loading optimization algorithm to schedule and adjust the loading capacity of each loading station at each time; for any day, add up the number of trains loaded at the loading station on that day to obtain the loading train number of the corresponding loading station; add up the loading train numbers of all loading stations to obtain the optimized coal train number for the corresponding day; add up the planned loading train numbers of all loading stations on that day to obtain the planned train number; Judge the number of coal trains out of the area and the planned number of trains out of the area; If equal, no adjustment; If there is a difference, call the coal train heavy-to-heavy back-to-empty optimization algorithm; For any path, the number of heavy-to-heavy trains on the path is accumulated for all heavy trains loaded at the loading station on the same day, and the heavy-to-heavy amount of the corresponding path is obtained; For any path, compare the heavy-to-heavy amount with the corresponding downstream heavy-to-heavy distribution plan; If the number is equal, no adjustment is made; If there is a deviation, call the railway downstream heavy-to-heavy distribution optimization algorithm.

3. The integrated coal supply chain scheduling decision method according to claim 1, characterized in that: The process of constructing a railway section capacity decline scenario decision model for algorithm combination calling is: For each railway path, the capacity of all times on the railway path on the same day is accumulated to obtain the capacity of the corresponding railway path on the same day; compare the capacity of the railway path on the same day with the planned capacity of the corresponding railway path on the same day; when the capacity of the railway path is not less than the planned capacity of the corresponding railway path, no adjustment is made; When the capacity of the railway path is less than the planned capacity of the corresponding railway path, call the coal train heavy-to-heavy back-to-empty optimization algorithm; Calculate the heavy-to-heavy amount of each path after calling the algorithm; for any path, compare the heavy-to-heavy amount with the corresponding downstream heavy-to-heavy distribution plan; If the number is equal, no adjustment is made; If there is a deviation, call the loading optimization algorithm and the railway downstream heavy-to-heavy distribution optimization algorithm to adjust the loading plan and the unloading distribution plan.

4. The integrated coal supply chain scheduling decision method according to claim 1, characterized in that: The process of constructing a key site heavy-to-heavy receiving and sending capacity decline scenario model for algorithm combination calling is: For each railway station, divide the number of tracks in maintenance by the total number of tracks of the corresponding railway station to obtain the heavy-to-heavy receiving and sending capacity decline degree of the corresponding railway station; for each railway path, multiply the railway station in maintenance and the correlation matrix of the path in turn by the heavy-to-heavy receiving and sending capacity decline degree of the corresponding station and the original capacity of the path to obtain the capacity deviation; subtract the corresponding capacity deviation from the original capacity of the railway path to obtain the capacity of the railway path during maintenance; When the capacity of the railway path during maintenance is less than the running plan, i.e. the capacity and train running plan deviation occurs, if so, call the coal train heavy-to-heavy back-to-empty optimization algorithm; When the capacity of the railway path is less than the planned capacity of the corresponding railway path, i.e. the actual capacity and train running plan deviation, call the coal train heavy-to-heavy back-to-empty optimization algorithm; calculate the heavy-to-heavy amount of each path after calling the algorithm; for any path, compare the heavy-to-heavy amount with the corresponding downstream heavy-to-heavy distribution plan; If the number is equal, no adjustment is made; If there is a deviation, call the loading optimization algorithm and the railway downstream heavy-to-heavy distribution optimization algorithm to adjust the loading plan and the unloading distribution plan.

5. The integrated coal supply chain scheduling decision method according to claim 1, characterized in that: The process of constructing a port unloading abnormal scenario decision model for algorithm combination calling is: The ratio of the number of dumper machines under maintenance to the number of dumper machines in the corresponding port is taken as the decline degree of heavy vehicle unloading capacity of the corresponding port front station; a maintenance schedule of the port is obtained to determine the maintenance condition of the port at each time; for any time, the decline degree of heavy vehicle unloading capacity of the port front station under maintenance at the time is multiplied by the predicted unloading amount of the port front station at the corresponding time to obtain an unloading amount difference; the predicted unloading amount of the port front station is subtracted by the unloading amount difference to obtain the unloading amount when the port front station is under maintenance; The unloading amount when the port front station is under maintenance is compared with the original port unloading plan amount; If the unloading amount when the port front station is under maintenance is not less than the original port unloading plan amount, no adjustment is made; If the unloading amount when the port front station is under maintenance is less than the original port unloading plan amount, a railway downstream heavy vehicle shunting optimization method is called to optimize the unloading shunting plan amount of the port; for each path, the actual heavy shunting train number after the railway downstream heavy vehicle shunting plan is called is calculated; The actual heavy shunting train number on each path is compared with the original heavy shunting train number; if they are the same, no adjustment is made; if they are different, a coal train heavy empty optimization algorithm and a loading optimization algorithm are called.

6. The integrated coal supply chain scheduling decision-making method according to claim 5, characterized in that: The process of calculating the actual heavy shunting train number after the railway downstream heavy vehicle shunting plan is called is as follows: for each time, the number of all trains unloaded at the port at the time is accumulated to obtain the unloading plan amount of the corresponding port at the corresponding time; for each path, the port unloading plan amount after the railway downstream heavy vehicle shunting plan is called is subtracted by the original port unloading plan amount, and then multiplied by the correlation coefficient of the port and the corresponding path to obtain the heavy shunting deviation of the corresponding path; the original heavy shunting train number on the path is added by the corresponding heavy shunting deviation to obtain the actual heavy shunting train number of the corresponding path.

7. The integrated coal supply chain scheduling decision-making method according to claim 1, characterized in that: The process of constructing a port marine closed navigation scenario decision-making model for algorithm combination calling is as follows: closed navigation time information is obtained, and a closed navigation early warning is determined to judge whether it is a marine closed navigation scenario; if it is a marine closed navigation scenario, a ship replacement optimization method and a port ship arrangement and shipping plan optimization algorithm are called; Based on the closed navigation time information, the port inventory at each time is calculated; when the port inventory does not exceed the upper limit of the port safety inventory, no adjustment is made; When the port inventory exceeds the upper limit of the port safety inventory, a railway downstream heavy vehicle shunting optimization algorithm is called; The actual heavy shunting train number after the railway downstream heavy vehicle shunting plan is called is calculated; the actual heavy shunting train number on each path is compared with the original heavy shunting train number; if they are the same, no adjustment is made; if they are different, a coal train heavy empty optimization algorithm and a loading optimization algorithm are called.

8. The integrated coal supply chain scheduling decision-making method according to claim 7, characterized in that: The process of calculating the port inventory at each time is as follows: ​ ​ For any time, multiply the planned port unloading plan quantity at that time by the average train load as the planned port unloading weight; for each berth, multiply the total number of ships that have been loaded at the berth and have not departed before the start of the closure by the load capacity of the corresponding ship as the numerator, and take the transshipment rate of the corresponding berth as the denominator to obtain the loading time of the berth before the end of the closure; add the loading time to the start time of the closure to obtain the end time of the loading at the corresponding berth at the time of the closure; for any time before the end of the loading or after the end of the closure, take the total loading quantity of the berth as the total loading quantity of the corresponding berth at that time; for any time, subtract the total loading quantity of all berths at that time from the planned port unloading weight to obtain the port inventory at the corresponding time.

9. The integrated coal supply chain scheduling decision-making method of claim 1, characterized in that: The process of constructing the power plant inventory emergency scenario decision-making model for algorithm combination calling is: For each power plant at any time, add up the load capacities of all ships that have arrived at the corresponding power plant and are associated with the power plant within the time step before the corresponding time as the unloading quantity of the corresponding power plant at the corresponding time; for each power plant, subtract the inventory of the corresponding power plant at the time step before the current time from the coal consumption within the time step, add the unloading quantity to obtain the inventory of the corresponding power plant at the current time; When the inventory of the corresponding power plant at the current time is not lower than the predetermined power plant safety inventory, no adjustment is made; When it is lower than the predetermined power plant safety inventory, subtract the predetermined power plant safety inventory from the initial inventory, divide by the coal consumption within the time step to calculate the latest ship arrival time length to avoid the situation of meeting the predetermined power plant safety inventory; When the latest ship arrival time length is greater than the total consumption time of the route of the power plant, the port ship scheduling and shipping plan optimization algorithm is called; if the latest ship arrival time length is not greater than the total consumption time of the route of the power plant, the ship replacement algorithm is called.

10. An integrated coal supply chain scheduling decision system using the method of any one of claims 1 to 9, characterized in that, It comprises: A basic scheduling optimization algorithm construction module that constructs multiple basic scheduling optimization algorithms covering coal supply chain railway transportation to port shipping; A coal loading quantity insufficient scenario algorithm calling module that calls multiple basic scheduling optimization algorithms in combination for different scenarios; when the predicted loading quantity of an upstream loading site is lower than the original loading plan, a coal loading quantity insufficient scenario decision-making model is constructed; by calculating the coal train number out of the zone, the loading train number deviation of the loading site, and the heavy train interchanging quantity, the coal train loading optimization, heavy train interchanging and empty train optimization, and heavy train shunting optimization algorithms downstream of the railway are called in combination; A railway section transport capacity reduction scenario algorithm calling module that calculates the transport capacity of the railway route on the same day based on the real-time transport capacity of each section of the railway line; when the transport capacity on the same day is less than the planned transport capacity, a railway section transport capacity reduction scenario decision-making model is constructed; by calculating the transport capacity of the railway route, the heavy train interchanging quantity, the coal train loading optimization, heavy train interchanging and empty train optimization, and heavy train shunting optimization algorithms downstream of the railway are called in combination; The key station heavy vehicle receiving and sending capacity decline scene algorithm calling module constructs a key station heavy vehicle receiving and sending capacity decline scene model when the receiving and unloading capacity of the station is lower than the predetermined receiving and unloading capacity; the railway path capacity during railway maintenance, heavy vehicle interchange weight, coal train interchange empty optimization, loading optimization, and railway downstream heavy vehicle shunting optimization algorithms are combined and called through calculation; The port unloading abnormal scene algorithm calling module constructs a port unloading abnormal scene decision model when the receiving and unloading capacity of the port is lower than the predetermined capacity; the railway downstream heavy vehicle shunting optimization, coal train interchange empty optimization, and loading optimization algorithms are combined and called through calculation of the port receiving and unloading capacity during maintenance, actual interchange heavy shunting train number, and actual interchange heavy shunting train number; The port sea shipping closure scene algorithm calling module constructs a port sea shipping closure scene decision model when the port prohibits ships from entering and leaving; the railway downstream heavy vehicle shunting optimization, coal train interchange empty optimization, and loading optimization algorithms are combined and called through calculation of the port inventory, actual interchange heavy shunting train number, and actual interchange heavy shunting train number; The power plant inventory emergency scene algorithm calling module constructs a power plant inventory emergency scene decision model when a ship fault causes the corresponding power plant to be unable to arrive on time or the inventory is lower than the predetermined power plant safety inventory; The port ship scheduling and shipping plan optimization, ship replacement algorithm are combined and called through calculation of the inventory of the downstream power plant, the longest ship arrival time, and the longest ship arrival time.

11. A terminal comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is used to operate according to the instructions to perform the steps of the method according to any one of claims 1-9.

12. A computer readable storage medium having stored thereon a computer program, characterized in that The program is executed by the processor to implement the steps of the method according to any one of claims 1-9.

Images