An intelligent scheduling method for space-time collaborative operations based on digital twin simulation

Abstract

The present invention provides an intelligent scheduling method for simulating spatiotemporal collaborative operations based on digital twins, which relates to the field of intelligent scheduling technology, including: S1: collecting combat-related data, building a digital twin model based on the combat-related data, and verifying the digital twin model, wherein the combat-related data includes scene data, environmental data, equipment parameters, and historical combat data; S2: dividing resource categories according to battlefield information, matching equipment for each resource category, and forming a resource group; S3: obtaining scheduling tasks, and using a multi-dimensional scheduling algorithm in the digital twin model to perform optimal scheduling according to the scheduling tasks and resource groups, and obtaining the optimal scheduling plan for the scheduling tasks. The present invention not only formulates a comprehensive and detailed scheduling plan based on the three core dimensions of tactical effect, resource utilization, and energy efficiency, but also can flexibly adjust and give corresponding optimization plans according to changes in combat needs.

Description

Technical Field

[0001] The present invention relates to the field of intelligent scheduling technology, and in particular to an intelligent scheduling method based on digital twin simulation of space-time collaborative operations. Background Art

[0002] Operational resource scheduling is a hot topic in the field of military command and control. Spatiotemporal collaborative operational scheduling refers to the use of various spatiotemporal information technology means to dynamically adjust and optimize operational plans in military operations to achieve collaborative operations among multiple combat units. It involves multiple aspects, including battlefield intelligence collection, tactical decision-making, resource scheduling, etc.

[0003] Time-space coordinated combat scheduling is an innovative combat scheduling method that emphasizes the efficient coordination and precise allocation of time and space resources during combat to optimize combat effectiveness. Time-space coordinated combat scheduling makes full use of time resources and space distance to form close cooperation and jointly complete combat tasks.

[0004] With the continuous evolution of modern warfare, space-time coordinated combat scheduling also needs to constantly adapt to new combat needs and technological developments. For example, through digital twin technology, we can simulate various possible combat scenarios and conditions, and quickly evaluate the cost-effectiveness of different scheduling strategies, thereby significantly improving scheduling efficiency. In addition, digital twin technology also has the ability to intelligently analyze the combat process, which can further optimize the space-time coordinated combat scheduling plan and push its level to a new height. The application of this technology will undoubtedly bring revolutionary changes to the space-time coordinated combat scheduling in modern warfare. Summary of the invention

[0005] In view of this, the present invention provides an intelligent scheduling method for space-time collaborative operations based on digital twin simulation. It not only formulates a comprehensive and detailed scheduling plan based on the three core dimensions of tactical effect, resource utilization, and energy efficiency, but also can flexibly adjust and give corresponding optimization plans according to changes in combat needs.

[0006] The technical purpose of the present invention is achieved in this way:

[0007] The present invention provides an intelligent scheduling method for space-time collaborative operations based on digital twin simulation, comprising the following steps:

[0008] S1: Collect combat-related data, build a digital twin model based on the combat-related data, and verify the digital twin model. The combat-related data includes scenario data, environmental data, equipment parameters, and historical combat data.

[0009] S2: Divide resource categories according to battlefield information, match equipment for each resource category, and form resource groups;

[0010] S3: Obtain the scheduling task, and use the multi-dimensional scheduling algorithm in the digital twin model to perform optimal scheduling based on the scheduling task and resource group to obtain the optimal scheduling plan for the scheduling task.

[0011] Based on the above technical solution, preferably, step S1 includes:

[0012] S11: Collect combat-related data, including scenario data, environmental data, equipment parameters, and historical combat data, and standardize combat-related data;

[0013] S12: Construct a three-dimensional digital model of the combat scenario based on combat-related data, including simulated terrain, buildings, weather systems, mobile combat units, and environmental conditions;

[0014] S13: setting variable parameters and configuring them into the three-dimensional digital model;

[0015] S14: Acquire actual data, and calibrate the three-dimensional digital model according to the actual data to adjust parameters of the three-dimensional digital model to obtain a digital twin model.

[0016] On the basis of the above technical scheme, preferably, the resource categories include communication, support, defense, attack and reconnaissance. The communication category is matched with communication equipment and communication systems, the support category is matched with transportation equipment and maintenance equipment, the defense category is matched with protection equipment and air defense missile systems, the attack category is matched with land and air attack equipment and missile systems, and the reconnaissance category is matched with reconnaissance equipment and radar equipment.

[0017] Based on the above technical solution, preferably, step S3 includes:

[0018] S31: Initialize the digital twin model and scheduling environment;

[0019] S32: Obtain a scheduled task X, define a target priority according to the scheduled task X, and construct an objective function F according to the target priority;

[0020] S33: construct a solution space according to the scheduling task X and the resource group R, the solution space includes task allocation and resource configuration scheme;

[0021] S34: Search for a solution set of the scheduling task X in the solution space using a multi-dimensional scheduling algorithm, evaluate the solution set, select a Pareto frontier solution from the solution set, and select an optimal solution from the Pareto frontier solution, that is, an optimal scheduling solution for the scheduling task X.

[0022] Based on the above technical solution, preferably, step S34 includes:

[0023] S341: setting an optimization target according to the scheduling task T, where the optimization target includes tactical effect, resource utilization and energy efficiency;

[0024] S342: Determine the number of individuals, initial pheromone evaporation rate, pheromone strength and heuristic factor, initialize pheromone concentration on all paths, and generate initial solutions using a greedy strategy;

[0025] S343: Each individual selects its path according to the pheromone concentration and the heuristic factor, performs resource scheduling simulation, and after completing the path, that is, completing a task scheduling, updates the pheromone concentration according to the quality of the path;

[0026] S344: Each individual constructs a solution set based on the selected path, evaluates the quality of the solution set based on the objective function, and updates the Pareto frontier solution;

[0027] S345: Adaptively adjust the pheromone evaporation rate according to the quality of the optimal solution and the average solution in the solution set, and update the pheromone concentration according to the Pareto frontier solution;

[0028] S346: Set a period t, and use the crossover and mutation method to generate a new solution set every other period t, and redistribute it to individuals;

[0029] S347: Repeat steps S343-S346 until the iteration stop condition is reached, and output the optimal solution in the Pareto frontier solution, that is, the optimal scheduling plan for scheduling task T.

[0030] Based on the above technical solution, preferably, the objective function is:

[0031]

[0032] Where F is the value of the objective function, n = 1, 2, 3, C n represents the target priority, f n (C n ) represents the dynamic weight, T n represents the nth optimization objective, P(T n ) indicates T n The penalty function of .

[0033] Based on the above technical solution, preferably, step S345 includes:

[0034] Step 1: Calculate the optimal solution quality Q in the current solution set best and the average solution quality Q of the entire solution set avg ;

[0035] Step 2: Calculate the optimal solution quality Q best and average solution quality Q avgThe mass difference ΔQ, ΔQ = Q best -Q avg ;

[0036] Step 3: Update the pheromone evaporation rate ρ according to the value of ΔQ. The pheromone evaporation rate ρ includes the global evaporation rate ρ global and the local evaporation rate ρ local ;

[0037] Step 4: Select a Pareto frontier solution from the current solution set, use the Pareto frontier solution to update the global pheromone concentration, and for each solution s on the Pareto frontier, update the pheromone concentration on the path (i, j);

[0038] Step 5: For non-Pareto solution paths, update the local pheromone concentration.

[0039] Based on the above technical solution, preferably, the updating formula of the pheromone evaporation rate is:

[0040] ρ type =ρ type，0 +k·tanh(γ·ΔQ)

[0041] In the formula, ρ type is the updated pheromone evaporation rate, where type indicates global or local, and ρ type，0 is the initial pheromone evaporation rate, k is the adjustment amplitude coefficient, γ is the adjustment sensitivity coefficient, and tanh is the hyperbolic tangent function, which is used to limit the adjustment amplitude to the range of [-k, k].

[0042] Based on the above technical solution, preferably, the update formula of the global pheromone concentration is:

[0043]

[0044] In the formula, τ ij represents the pheromone concentration, ρ global is the global evaporation rate, S Pareto is the Pareto frontier solution, F -1 (s) represents the inverse of the objective function of solution s, Z is the normalization factor, represents the pheromone increment of solution s;

[0045] The update formula of local pheromone concentration is:

[0046] τ ij =(1-ρ local α ij )·τ ij

[0047] In the formula, α ijis the dynamic coefficient of path (i, j).

[0048] Based on the above technical solution, preferably, the heuristic factor is expressed as:

[0049]

[0050] Where η ij is the heuristic factor, d y (i, j) represents the cost of path (i, j) on the yth optimization target, w x is the weight of the xth optimization objective, U xy Represents the correlation matrix between the xth optimization objective and the yth optimization objective.

[0051] The method of the present invention has the following beneficial effects compared with the prior art:

[0052] (1) The present invention realizes dynamic combat environment simulation and parameter adjustment by constructing a high-precision digital twin model synchronized with actual data. The method optimizes resource allocation and improves utilization efficiency by clearly dividing resource categories and matching appropriate equipment. The multi-dimensional scheduling algorithm is applied, combined with multi-objective optimization, so that the scheduling scheme reaches the optimal level while meeting multiple priorities and constraints. In addition, the Pareto frontier solution selection strategy is adopted to ensure the best balance in multi-objective decision-making, effectively improve the accuracy and adaptability of combat scheduling, and enhance the ability to respond to complex and changing combat environments;

[0053] (2) The present invention effectively handles the complex decision-making problems in combat scheduling by defining a multi-objective optimization function in detail. By optimizing the objective function F, the method not only takes into account tactical effects, resource utilization, and energy efficiency, but also introduces dynamic weights and penalty functions, so that the scheduling strategy can be flexibly adjusted to adapt to different combat needs and conditions. This method can minimize resource consumption while ensuring combat efficiency, thereby achieving an optimal balance between multiple key performance indicators;

[0054] (3) The pheromone management and update mechanism of the present invention is the key to achieving effective path selection and resource scheduling. The initial solution generated by the initial pheromone concentration setting and the greedy strategy provides a good starting point for the algorithm. Subsequently, by adaptively adjusting the pheromone evaporation rate and updating the pheromone concentration according to the Pareto frontier solution, the method can dynamically adjust the search strategy to reflect the latest scheduling priorities and environmental changes. This intelligent pheromone management not only improves the convergence speed of the algorithm, but also enhances the diversity and quality of the solution;

[0055] (4) The present invention generates a new solution set by adopting the crossover and mutation method, and uses the multi-dimensional scheduling algorithm for iterative optimization, which significantly improves the efficiency and quality of solving the scheduling task. In each iteration cycle, by evaluating the quality of the solution set and selecting the Pareto frontier solution, the method ensures the continuous optimization and adaptive adjustment of the solution. This iterative optimization process not only accelerates the search for the optimal solution, but also responds to the complex and changing combat environment by carefully adjusting the scheduling strategy, and finally outputs the optimal scheduling solution to ensure the successful execution of the combat mission. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0057] Figure 1 The figure is a flow chart of a method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0058] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0059] like Figure 1 As shown, the present invention provides an intelligent scheduling method for space-time collaborative operations based on digital twin simulation, comprising the following steps:

[0060] S1: Collect combat-related data, build a digital twin model based on the combat-related data, and verify the digital twin model. The combat-related data includes scenario data, environmental data, equipment parameters, and historical combat data.

[0061] S2: Divide resource categories according to battlefield information, match equipment to each resource category, and form resource groups, where equipment includes test scores;

[0062] S3: Obtain the scheduling task, and use the multi-dimensional scheduling algorithm in the digital twin model to perform optimal scheduling based on the scheduling task and resource group to obtain the optimal scheduling plan for the scheduling task.

[0063] Specifically, in one embodiment of the present invention, step S1 includes:

[0064] S11: Collect combat-related data, including scenario data, environmental data, equipment parameters, and historical combat data, and standardize combat-related data.

[0065] First, collect all kinds of data directly related to the combat mission. This includes but is not limited to scenario data, such as geographic location, terrain features, environmental data, such as meteorological conditions, temperature, humidity, equipment parameters, such as performance indicators of weapon systems, vehicle mobility, and historical combat data, previous combat experience, and records of enemy and friendly actions. The collected data needs to be standardized to ensure the consistency of the data format and data quality, so as to facilitate subsequent analysis and simulation. Standardization includes data cleaning to remove invalid and erroneous data points, data conversion, such as unified measurement units, time formats, etc., and data encoding, such as the digitization of classified data.

[0066] S12: Construct a three-dimensional digital model of the combat scenario based on combat-related data, including simulated terrain, buildings, weather systems, mobile combat units, and environmental conditions.

[0067] A three-dimensional digital model is constructed using the standardized data. This model is a virtual representation of the combat environment, including the ups and downs of the terrain, the layout of buildings, and the dynamic changes of the weather system. In addition, the model will also simulate mobile combat units, such as tanks, aircraft, soldiers, etc., and the impact of various environmental conditions on combat.

[0068] S13: Setting variable parameters and configuring them into the three-dimensional digital model.

[0069] Variable parameters can be set in the 3D model to simulate different combat scenarios. For example, weather conditions such as rainfall and wind speed can be adjusted, or the deployment location and number of certain combat units can be changed.

[0070] S14: Acquire actual data, and calibrate the three-dimensional digital model according to the actual data to adjust parameters of the three-dimensional digital model to obtain a digital twin model.

[0071] During the simulation process, actual combat data needs to be obtained in real time. These data are used to correct and adjust the parameters of the three-dimensional digital model to ensure the accuracy and real-time performance of the model.

[0072] Let's take a specific example to illustrate:

[0073] This embodiment is a simulation process of maritime combat. First, detailed data on the geographical location of the ocean needs to be collected, including the depth, tides and currents of the sea area. Meteorological condition data, such as wind speed, wind direction, wave height, temperature and humidity, etc., are collected. The performance indicators of various types of ships and submarines, such as speed, durability, range and accuracy of weapon systems, etc., are recorded in detail. Past naval battle records are analyzed, including the use of tactics, action records of both sides and the results of the battle.

[0074] Next, remove erroneous and invalid data points, such as old data that does not meet current technical standards. Unify the measurement units of the data (such as converting nautical miles to kilometers), time format, etc. Numerical processing of classified data is performed to facilitate subsequent analysis and simulation.

[0075] Use Unity to build a detailed 3D model of the sea terrain, weather system, and ships. Use the physics engine to simulate the impact of waves and wind on ships, as well as the dynamic interaction between ships. Integrate a dynamic weather system to simulate various weather conditions from calm to stormy. Set the physical properties of the ship, such as buoyancy, speed, maneuverability, etc., based on the collected data.

[0076] Configure variable parameters for the 3D digital model to adjust weather parameters (wind speed, rainfall), currents, and tides according to tactical requirements. Configure the deployment of combat units, such as the formation of ships, the lurking depth and position of submarines. Conduct tactical simulations with different parameter settings to test combat effectiveness under various environmental conditions. Evaluate the impact of different tactical options on combat results, such as assault, retreat, or defense.

[0077] During the simulation process, the data from the simulation training, such as the actual movement speed of the ship and the operation data of the weapon system, are accessed in real time. The parameters of the 3D model are adjusted and corrected based on the feedback from the actual operation.

[0078] Specifically, in one embodiment of the present invention, resource categories include communication, support, defense, attack and reconnaissance. The communication category is matched with communication equipment and communication systems, the support category is matched with transportation equipment and maintenance equipment, the defense category is matched with protection equipment and air defense missile systems, the attack category is matched with land and air attack equipment and missile systems, and the reconnaissance category is matched with reconnaissance equipment and radar equipment.

[0079] Among them, the communication category usually needs to be equipped with communication equipment and communication systems. Communication equipment: including handheld radios, vehicle-mounted radios, portable communication equipment, etc., are mainly used to achieve instant communication on the battlefield to ensure that commanders and combat units can exchange information quickly and effectively. Communication system: Advanced communication systems such as satellite communication systems, optical fiber communication networks, encrypted communication systems, etc. are used to provide long-distance and secure communication links to support complex command and control operations and strategic-level communications.

[0080] Support equipment needs to be equipped with transportation equipment and maintenance equipment. Transportation equipment: including military trucks, armored transport vehicles, ships and transport aircraft, etc., used to transport troops, weapons, ammunition and other materials to ensure the supply and rapid deployment of front-line troops. Maintenance equipment: including engineering vehicles, maintenance tools and facilities, used to maintain and repair weapon systems, vehicles and other military equipment to maintain the continuity of combat effectiveness and the operating efficiency of equipment.

[0081] Defense requires protective equipment and anti-aircraft missile systems. Protective equipment: including protective walls, bunkers, bulletproof vests and other protective structures used to protect soldiers and key facilities from enemy attacks. Anti-aircraft missile systems: high-tech air defense systems, such as surface-to-air missiles and anti-aircraft guns, are used to detect, track and intercept incoming flying objects such as airplanes, helicopters and missiles to protect important targets and troops.

[0082] The attack category needs to be equipped with land and air attack equipment and missile systems. Land and air attack equipment: including tanks, self-propelled artillery, fighter jets and attack helicopters, used to carry out effective ground and air attacks on the enemy and destroy the enemy's combat power and strategic resources. Missile systems: including surface-to-surface missiles, cruise missiles and tactical ballistic missiles, provide precise long-range strike capabilities, used to destroy the enemy's command center, communication nodes and other key targets.

[0083] Reconnaissance equipment needs to be equipped with reconnaissance equipment and radar equipment. Reconnaissance equipment: including unmanned reconnaissance aircraft (UAV), reconnaissance vehicles and special reconnaissance equipment, used to collect intelligence on the enemy's position, activities and resources, and provide support for combat deployment and tactical decision-making. Radar equipment: including ground radar, shipborne radar and air radar systems, used to monitor the dynamics of the air, land and sea in real time, quickly identify enemy units and equipment, and provide intelligence support and target indication.

[0084] Specifically, in one embodiment of the present invention, step S3 includes:

[0085] S31: Initialize the digital twin model and scheduling environment;

[0086] S32: Obtain a scheduled task X, define a target priority according to the scheduled task X, and construct an objective function F according to the target priority;

[0087] S33: construct a solution space according to the scheduling task X and the resource group R, the solution space includes task allocation and resource configuration scheme;

[0088] S34: Search for a solution set of the scheduling task X in the solution space using a multi-dimensional scheduling algorithm, evaluate the solution set, select a Pareto frontier solution from the solution set, and select an optimal solution from the Pareto frontier solution, that is, an optimal scheduling solution for the scheduling task X.

[0089] Specifically, after the digital twin model is constructed, the scheduling environment is initialized according to its corrected parameters, that is, in the digital twin model, adjustable values ​​are set to cope with different scheduling tasks or combat strategies.

[0090] Obtain specific dispatching tasks, such as maritime defense, maritime strike, siege attack, etc. According to the nature of each dispatching task and the battlefield situation, determine the priority of the optimization target. Based on this, the objective function constructed in this embodiment is as follows:

[0091]

[0092] Where F is the value of the objective function, n = 1, 2, 3, C n represents the target priority, f n (C n ) represents the dynamic weight, T n represents the nth optimization objective, P(T n ) indicates T n The penalty function of .

[0093] The optimization goals in this embodiment include maximizing tactical effects, maximizing resource utilization, and maximizing energy efficiency. Considering that during combat, as the task becomes tight or the external environment changes, different target priorities should dynamically adjust the optimization weights of different optimization goals. Therefore, a dynamic weight function is used to adjust the weights based on the target priority. At the same time, since there will be abnormal situations such as effect fluctuations, low utilization, and rapid energy bottoming out, a penalty function is set to avoid this.

[0094] According to the scheduling tasks and resource groups, a solution space containing all possible task allocation and resource configuration schemes is constructed. Specifically:

[0095] Detailed analysis of available resources and their characteristics, such as equipment performance, location, etc., each equipment including its test score, and task requirements. Create a solution space that includes all possible resource configurations and task allocation schemes, each of which is a potential scheduling solution.

[0096] Apply a multi-dimensional scheduling algorithm to search for the optimal solution in the solution space, and select the scheduling solution that best suits the current operational needs. Step S34 includes:

[0097] S341: setting an optimization target according to the scheduling task T, where the optimization target includes tactical effect, resource utilization and energy efficiency;

[0098] S342: Determine the number of individuals, initial pheromone evaporation rate, pheromone strength and heuristic factor, initialize pheromone concentration on all paths, and generate initial solutions using a greedy strategy;

[0099] S343: Each individual selects its path according to the pheromone concentration and the heuristic factor, performs resource scheduling simulation, and after completing the path, that is, completing a task scheduling, updates the pheromone concentration according to the quality of the path;

[0100] S344: Each individual constructs a solution set based on the selected path, evaluates the quality of the solution set based on the objective function, and updates the Pareto frontier solution;

[0101] S345: Adaptively adjust the pheromone evaporation rate according to the quality of the optimal solution and the average solution in the solution set, and update the pheromone concentration according to the Pareto frontier solution;

[0102] S346: Set a period t, and use the crossover and mutation method to generate a new solution set every other period t, and redistribute it to individuals;

[0103] S347: Repeat steps S343-S346 until the iteration stop condition is reached, and output the optimal solution in the Pareto frontier solution, that is, the optimal scheduling plan for scheduling task T.

[0104] Let's use a specific example to illustrate:

[0105] According to the scheduling task T, the optimization objectives are determined, including tactical effect, resource utilization and energy efficiency.

[0106] Determine the number of individuals and set the initial global evaporation rate ρ global =0.1, local evaporation rate ρ local =0.01.

[0107] Initialize the pheromone intensity on all paths to 0.5. Calculate the heuristic factor for each path. Generate the initial solution using the greedy strategy and select the path with the highest heuristic factor as the starting solution.

[0108] The calculation formula of the heuristic factor is:

[0109]

[0110] Where η ij is the heuristic factor, d y (i,j) represents the cost of path (i,j) on the yth optimization target, w x is the weight of the xth optimization objective, U xy Represents the correlation matrix between the xth optimization objective and the yth optimization objective.

[0111] In this embodiment, considering multiple optimization objectives (such as tactical effect, resource utilization and energy efficiency), y(i,j) represents the cost of path (i,j) on the yth optimization objective. This can more comprehensively evaluate the pros and cons of a path and help find a solution that performs well on multiple objectives. By introducing an inter-objective correlation matrix U xy , to evaluate the comprehensive quality of a path.

[0112] Each individual selects a path in a roulette wheel based on the current pheromone concentration and heuristic factor to perform a complete scheduling task, that is, a scheduling plan from the starting point to the end point. After completing the task, the pheromone concentration on the path is updated based on the actual performance of the path (such as tactical effect, resource utilization, and energy efficiency). The pheromone concentration increases more for paths with good performance; otherwise, the increase is small.

[0113] A complete scheduling path obtained for each individual is actually a feasible solution. The solutions of all individuals are formed into a solution set. Then, the quality of each solution is evaluated using the preset objective function to obtain the objective function value of each solution in the solution set. Next, the Pareto frontier solutions are selected from the solution set, that is, those solutions that perform best on any objective and are not completely dominated by other solutions, as the current Pareto frontier solutions.

[0114] Update the Pareto frontier solution set to ensure that non-inferior solutions are stored:

[0115] Compare the new solution to the existing Pareto solutions.

[0116] If the new solution dominates any existing solution, the dominated solution is replaced.

[0117] If the new solution is non-inferior and not dominated by any existing solution, it is added to the Pareto set.

[0118] In order to balance the exploration and utilization of the algorithm, the pheromone evaporation rate needs to be adjusted dynamically. The specific steps are as follows:

[0119] Step 1: Calculate the optimal solution quality Q in the current solution set best and the average solution quality Q of the entire solution set avg .

[0120] Specifically, in this embodiment, the quality Q is a scalar value of the objective function value. Traverse all solutions in the current solution set and find the solution with the highest quality. Its quality is Q best At the same time, the average quality value of all solutions is calculated as Q avg .

[0121] Step 2: Calculate the optimal solution quality Q best and average solution quality Q avg The mass difference ΔQ, ΔQ = Q best -Qavg ; ΔQ reflects the quality gap between the current optimal solution and the overall solution. The larger the ΔQ, the greater the quality gap between the optimal solution and other solutions, and the algorithm may have converged; the smaller the ΔQ, the closer the quality of the solutions in the solution set is, and the algorithm may still be in the exploration stage.

[0122] Step 3: Update the pheromone evaporation rate ρ according to the value of ΔQ. The pheromone evaporation rate ρ includes the global evaporation rate ρ global and the local evaporation rate ρ local ; The pheromone evaporation rate ρ controls the speed of pheromone decay, which affects the balance between the global search ability and the local search ability of the algorithm. When ΔQ is large, it means that the algorithm has found a better solution. At this time, the pheromone decay should be accelerated to encourage the algorithm to conduct in-depth search near the current better solution; when ΔQ is small, it means that the algorithm has not found a satisfactory solution. At this time, the pheromone decay should be slowed down to encourage the algorithm to conduct more extensive exploration.

[0123] The updated formula for the pheromone evaporation rate is:

[0124] ρ type =ρ type，0 +k·tanh(γ·ΔQ)

[0125] In the formula, ρ type is the updated pheromone evaporation rate, where type indicates global or local, and ρ type，0 is the initial pheromone evaporation rate, k is the adjustment amplitude coefficient, γ is the adjustment sensitivity coefficient, and tanh is the hyperbolic tangent function, which is used to limit the adjustment amplitude to the range of [-k, k].

[0126] In this embodiment, when updating the pheromone evaporation rate, the hyperbolic tangent function tanh is used to map ΔQ to the interval [-k, k], and the evaporation rate can be adaptively adjusted. When ΔQ is large, tanh(γ·ΔQ) approaches 1, and the evaporation rate increases by about k; when ΔQ is small, tanh(γ·ΔQ) approaches 0, and the evaporation rate remains at the initial value ρ type，0 The introduction of adjustment amplitude coefficient k and adjustment sensitivity coefficient γ can flexibly control the adjustment amplitude of the evaporation rate and the sensitivity to the change of ΔQ. By adjusting these two parameters, the evaporation rate can be finely controlled at different stages of the algorithm.

[0127] Step 4: Select the Pareto frontier solution from the current solution set, and use the Pareto frontier solution to update the global pheromone concentration. For each solution s on the Pareto frontier, update the pheromone concentration on the path (i, j); the update formula for the global pheromone concentration is:

[0128]

[0129] In the formula, τ ij represents the pheromone concentration, ρ global is the global evaporation rate, S Pareto is the Pareto frontier solution, F -1 (s) represents the inverse of the objective function of solution s, Z is the normalization factor, represents the pheromone increment of solution s;

[0130] In this embodiment, the Pareto frontier solution is used to update the pheromone so that the algorithm searches in the direction of the non-dominated solution set, which helps to find more Pareto optimal solutions. The inverse of the objective function value of the solution is taken and then normalized so that the solution with a large objective function value can obtain more pheromone rewards, thereby providing more guidance for subsequent searches. The normalization factor Z is introduced to balance the contribution of different solutions to pheromones and avoid individual extreme solutions dominating the search direction.

[0131] Step 5: For paths that are not Pareto solutions, update the local pheromone concentration. For those solutions that are not on the Pareto frontier, the paths they involve also need to be updated with pheromones, but the update amplitude should be small to avoid excessively affecting the search direction. The update formula for the local pheromone concentration is:

[0132] τ ij =(1-ρ local α ij )·τ ij

[0133] In the formula, α ij is the dynamic coefficient of path (i, j).

[0134] In this embodiment, pheromone updates are performed on the paths involved in non-Pareto solutions to maintain the exploratory nature of the algorithm and avoid premature convergence. ij , dynamically adjust the pheromone evaporation rate according to the historical performance of path (i, j). For example, based on the frequency and quality of path (i, j) in the historical solution, calculate an α between 0 and 1 ij Value, α ij The larger it is, the better the historical performance of path (i, j) is, and the slower its pheromone should evaporate.

[0135] In this embodiment, according to the above five steps, in each round of iteration, the pheromone update strategy can be adaptively adjusted according to the quality of the current solution set, and a balance can be struck between exploration and utilization, so as to more efficiently search for a high-quality Pareto optimal solution set.

[0136] After every t iterations, the crossover and mutation operations of the genetic algorithm are used to generate a new solution set. That is, some individuals are randomly selected, and their paths are randomly crossovered or mutated to generate new paths. This can expand the search range of the solution space and prevent the algorithm from falling into the local optimum early.

[0137] Repeat the above steps until the preset iteration stop condition is reached, such as the number of iterations reaches the upper limit, or the quality of the solution is not significantly improved. Finally, the optimal solution is selected from the Pareto frontier solution and output as the optimal scheduling solution for scheduling task T.

[0138] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent scheduling method for space-time collaborative operations based on digital twin simulation, characterized in that: The following steps are involved: S1: Collect combat-related data, build a digital twin model based on the combat-related data, and verify the digital twin model. The combat-related data includes scenario data, environmental data, equipment parameters, and historical combat data. S2: Divide resource categories according to battlefield information, match equipment for each resource category, and form resource groups; S3: Obtain the scheduling task, and use the multi-dimensional scheduling algorithm to perform optimal scheduling in the digital twin model according to the scheduling task and resource group to obtain the optimal scheduling solution for the scheduling task; Step S3 includes: S31: Initialize the digital twin model and scheduling environment; S32: Obtain a scheduled task X, define a target priority according to the scheduled task X, and construct an objective function F according to the target priority; S33: construct a solution space according to the scheduling task X and the resource group R, the solution space includes task allocation and resource configuration scheme; S34: searching for a solution set of the scheduling task X in the solution space by using a multi-dimensional scheduling algorithm, evaluating the solution set, selecting a Pareto frontier solution from the solution set, and selecting an optimal solution from the Pareto frontier solution, that is, an optimal scheduling solution for the scheduling task X; Step S34 includes: S341: Setting an optimization target according to the scheduling task T, the optimization target including total time, total moving distance and total energy consumption; S342: Determine the number of individuals, initial pheromone evaporation rate, pheromone strength and heuristic factor, initialize pheromone concentration on all paths, and generate initial solutions using a greedy strategy; S343: Each individual selects its path according to the pheromone concentration and the heuristic factor, performs resource scheduling simulation, and after completing the path, that is, completing a task scheduling, updates the pheromone concentration according to the quality of the path; S344: Each individual constructs a solution set based on the selected path, evaluates the quality of the solution set based on the objective function, and updates the Pareto frontier solution; S345: Adaptively adjust the pheromone evaporation rate according to the quality of the optimal solution and the average solution in the solution set, and update the pheromone concentration according to the Pareto frontier solution; S346: Set a period t, and use the crossover and mutation method to generate a new solution set every other period t, and redistribute it to individuals; S347: Repeat steps S343-S346 until the iteration stop condition is reached, and output the optimal solution in the Pareto frontier solution, that is, the optimal scheduling solution for the scheduling task T; The objective function is: ； Where F is the value of the objective function, n=1,2,3, C n represents the target priority, f n (C n ) represents the dynamic weight, T n represents the nth optimization objective, P(T n ) indicates T n The penalty function of Step S345 includes: Step 1: Calculate the optimal solution quality Q in the current solution set best and the average solution quality Q of the entire solution set avg ; Step 2: Calculate the optimal solution quality Q best and average solution quality Q avg Quality Difference , ; Step 3: According to Update the value of pheromone evaporation rate , pheromone evaporation rate Including global evaporation rate and local evaporation rate ; Step 4: Select the Pareto frontier solution from the current solution set, use the Pareto frontier solution to update the global pheromone concentration, and for each solution s on the Pareto frontier, update the pheromone concentration on the path (i, j); Step 5: For non-Pareto solution paths, update the local pheromone concentration.

2. The intelligent scheduling method for space-time collaborative operations based on digital twin simulation according to claim 1 is characterized in that: Step S1 includes: S11: Collect combat-related data, including scenario data, environmental data, equipment parameters, and historical combat data, and standardize combat-related data; S12: Construct a three-dimensional digital model of the combat scenario based on combat-related data, including simulated terrain, buildings, weather systems, mobile combat units, and environmental conditions; S13: setting variable parameters and configuring them into the three-dimensional digital model; S14: Acquire actual data, and calibrate the three-dimensional digital model according to the actual data to adjust parameters of the three-dimensional digital model to obtain a digital twin model.

3. The intelligent scheduling method for space-time collaborative operations based on digital twin simulation as claimed in claim 2 is characterized in that: Resource categories include positioning, engagement, detection, indication and tracking. The positioning category is matched with sonar equipment and radar equipment, the engagement category is matched with missile equipment and torpedo equipment, the detection category is matched with sonar equipment and radar equipment, the indication category is matched with fire control systems and command centers, and the tracking category is matched with unmanned equipment and ships.

4. The intelligent scheduling method for space-time collaborative operations based on digital twin simulation according to claim 1 is characterized in that: The updated formula for the pheromone evaporation rate is: ； In the formula, is the updated pheromone evaporation rate, where type indicates global or local, is the initial pheromone evaporation rate, k is the adjustment amplitude coefficient, is the adjustment sensitivity coefficient, and tanh is the hyperbolic tangent function, which is used to limit the adjustment amplitude to the range of [-k, k].

5. The intelligent scheduling method for space-time collaborative operations based on digital twin simulation according to claim 1 is characterized in that: The update formula of global pheromone concentration is: ； In the formula, express t The global pheromone concentration at time +1, for t The global pheromone concentration at time , is the global evaporation rate, is the Pareto frontier solution, represents the inverse of the objective function of solution s, Z is the normalization factor, represents the pheromone increment of solution s; The update formula of local pheromone concentration is: ； In the formula, is the dynamic coefficient of path (i,j), for t The local pheromone concentration at time +1, for t The local pheromone concentration at the time.

6. The intelligent scheduling method for space-time collaborative operations based on digital twin simulation according to claim 1 is characterized in that: The heuristic factor is expressed as: ； In the formula, is the heuristic factor, represents the cost of path (i, j) on the yth optimization target, is the weight of the xth optimization objective, Represents the correlation matrix between the xth optimization objective and the yth optimization objective.

Images