Prefabricated component distribution scheduling optimization method based on time window and multiple vehicle types

Abstract

The application provides a prefabricated component distribution scheduling optimization method based on a time window and multiple vehicle types, and relates to the technical field of distribution scheduling. Parameter information of multiple transport vehicles and distribution time windows of construction sites are obtained and utilized. An optimization function with the dual objectives of minimizing distribution cost and maximizing vehicle average loading rate is constructed. By simultaneously optimizing the two objectives, a balanced deployment scheme between economy and resource utilization efficiency can be obtained, avoiding the one-sidedness that may be caused by single-objective optimization. A multi-objective ant colony-genetic hybrid optimization algorithm is used for solving, generating multiple deployment schemes with different trade-offs for users to select. Through comprehensive consideration of multiple vehicle types, time windows, dual-objective optimization and hybrid intelligent algorithms, an efficient and flexible prefabricated component distribution scheduling optimization solution is provided, effectively improving resource utilization efficiency in the multiple vehicle type mixed distribution scene and improving the practicality of the distribution scheme.

Description

Optimization Method for Prefabricated Component Delivery Scheduling Based on Time Windows and Multiple Vehicle Models Technical Field

[0001] This invention relates to the field of delivery scheduling technology, and more specifically, to a method for optimizing the delivery scheduling of prefabricated components based on time windows and multiple vehicle types. Background Technology

[0002] Prefabricated construction is a new type of green construction method. Compared with cast-in-place concrete construction, prefabricated construction has many advantages such as high efficiency, short construction period, controllable quality, environmental protection, and labor saving, and has therefore become a trend in the construction industry.

[0003] However, the high cost of prefabricated construction is a significant factor currently limiting the development of the industry. This cost primarily involves six stages: design, production, transportation, construction, operation and maintenance, and the market. In the transportation stage, the production plant delivers prefabricated components to the customer's location via vehicles.

[0004] In existing research on prefabricated component distribution optimization, most optimization models only consider a single delivery vehicle type. The loading rate of a single vehicle type is relatively fixed. However, in reality, multiple vehicle types are usually used for mixed transportation. As a result, in the scenario of mixed delivery of multiple vehicle types, the applicability and solution efficiency of optimization models that only consider a single delivery vehicle type are difficult to meet the actual engineering practice requirements. The actual utilization rate of multiple vehicle types is poor, and the delivery scheduling scheme needs to be optimized. Summary of the Invention

[0005] The problem that this invention aims to solve is that existing optimization models that only consider a single delivery vehicle type are insufficient in terms of applicability and solution efficiency to meet the needs of practical engineering in multi-vehicle mixed delivery scenarios.

[0006] To address the aforementioned problems, in a first aspect, the present invention provides a method for optimizing the delivery scheduling of prefabricated components based on time windows and multiple vehicle models, comprising:

[0007] Obtain parameter information of transport vehicles and construction sites, wherein the parameter information of transport vehicles includes the type of transport vehicle, and the parameter information of construction sites includes the delivery time window of the construction site.

[0008] Based on the parameter information of the transport vehicles and the construction site, a bi-objective optimization function is constructed with the optimization objectives of minimizing delivery costs and maximizing the average vehicle loading rate. The delivery cost consists of the fixed vehicle usage cost, transportation cost and penalty cost, and the average vehicle loading rate is the sum of the single loading rates of all trips divided by the total number of trips.

[0009] A multi-objective ant colony-genetic hybrid optimization algorithm is used to solve the bi-objective optimization function. Under the conditions of meeting the delivery time window requirements of the construction site and the preset constraints, multiple dispatch schemes for various types of transport vehicles are obtained. The multi-objective ant colony-genetic hybrid optimization algorithm is transformed into a genetic algorithm for a local search after a preset number of local searches using the multi-objective ant colony optimization algorithm.

[0010] Based on user preferences, a distribution plan is selected, and prefabricated components are delivered.

[0011] This invention provides a method for optimizing the delivery scheduling of prefabricated components based on time windows and multiple vehicle models. Compared with existing technologies, it has the following advantages:

[0012] The algorithm acquires and utilizes parameter information from various transport vehicles, including vehicle type, load capacity, and delivery time windows at construction sites. This allows the optimization model to fully consider the characteristics of different vehicles and the specific time requirements of construction sites, thus more closely reflecting actual delivery scenarios. Secondly, an optimization function is constructed with dual objectives: minimizing delivery costs and maximizing the average vehicle loading rate. The delivery cost composition considers fixed vehicle usage costs, transportation costs, and penalty costs, comprehensively reflecting the economic expenditures during the delivery process. Simultaneously, maximizing the average vehicle loading rate directly addresses the problem of poor utilization of multiple vehicle types in existing technologies. By simultaneously optimizing these two objectives, a balanced allocation scheme between economy and resource utilization efficiency can be obtained, avoiding the one-sidedness that may result from single-objective optimization. A multi-objective ant colony-genetic hybrid optimization algorithm is employed for solving the problem, combining the global search capability of the ant colony algorithm with the local fine-grained search capability of the genetic algorithm, enabling more effective searching for non-dominated solution sets in a complex solution space. Ant colony optimization's iterative search helps explore a broad solution space, while the introduction of genetic algorithms further optimizes and maintains diversity in the solution set after a predetermined number of local searches. This prevents the algorithm from prematurely converging to a local optimum, increasing the likelihood of finding a high-quality allocation solution. It can handle large-scale, highly complex delivery scheduling problems and generate multiple allocation solutions with different trade-offs for users to choose from. Users can flexibly select based on their actual operational needs, making the delivery scheduling solution more adaptable and practical. This method, by comprehensively considering multiple vehicle types, time windows, dual-objective optimization, and hybrid intelligent algorithms, provides a more comprehensive, efficient, and flexible prefabricated component delivery scheduling optimization solution. It effectively improves resource utilization efficiency in multi-vehicle mixed delivery scenarios, reduces delivery costs, and enhances the practicality of the delivery solution. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 is a flowchart illustrating a prefabricated component delivery scheduling optimization method based on time windows and multiple vehicle models provided in an embodiment of the present invention;

[0015] Figure 2 is a flowchart of the improved multi-objective ant colony-genetic hybrid optimization algorithm provided in an embodiment of the present invention;

[0016] Figure 3 is a distribution diagram of the optimal solution of the improved algorithm provided in the embodiment of the present invention;

[0017] Figure 4 is a distribution diagram of the optimal solution of the original algorithm provided in the embodiment of the present invention;

[0018] Figure 5 is a schematic diagram of a prefabricated component delivery scheduling optimization system based on time windows and multiple vehicle models provided in an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions in the embodiments of this application are described clearly and completely. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0021] As shown in Figure 1, an embodiment of this application provides a prefabricated component delivery scheduling optimization method based on time windows and multiple vehicle models, including:

[0022] S1: Obtain parameter information of transport vehicles and construction site, wherein the parameter information of transport vehicles includes the type of transport vehicle, and the parameter information of construction site includes the delivery time window of construction site.

[0023] S2: Based on the parameter information of the transport vehicles and the parameter information of the construction site, a bi-objective optimization function is constructed with the optimization objectives of minimizing delivery costs and maximizing the average vehicle loading rate. The delivery cost consists of the fixed usage cost of the vehicle, the transportation cost, and the penalty cost. The average vehicle loading rate is the sum of the single loading rates of all trips divided by the total number of trips.

[0024] S3: A multi-objective ant colony-genetic hybrid optimization algorithm is used to solve the bi-objective optimization function. Under the conditions of meeting the delivery time window requirements of the construction site and the preset constraints, multiple dispatch schemes for various types of transport vehicles are obtained. Among them, the multi-objective ant colony-genetic hybrid optimization algorithm is transformed into a genetic algorithm for a local search after a preset number of local searches using the multi-objective ant colony optimization algorithm.

[0025] S4: Select the allocation plan based on user preferences and deliver the prefabricated components.

[0026] In existing research on prefabricated component delivery optimization, most optimization models only consider a single delivery vehicle type. In actual multi-vehicle mixed transportation scenarios, this leads to limited applicability and solution efficiency of the optimization models, poor actual utilization of multiple vehicle types, and delivery scheduling schemes need to be optimized.

[0027] In this optional embodiment, parameter information of various transport vehicles is acquired and utilized, including vehicle type, load capacity, and delivery time windows at construction sites. This allows the optimization model to fully consider the characteristics of different vehicles and the specific time requirements of construction sites, thus more closely reflecting actual delivery scenarios. Secondly, an optimization function is constructed with dual objectives: minimizing delivery costs and maximizing the average vehicle loading rate. The delivery cost composition considers fixed vehicle usage costs, transportation costs, and penalty costs, comprehensively reflecting the economic expenditures during the delivery process. Simultaneously, maximizing the average vehicle loading rate directly addresses the problem of poor utilization of multiple vehicle types in existing technologies. By simultaneously optimizing these two objectives, a balanced allocation scheme between economy and resource utilization efficiency can be obtained, avoiding the one-sidedness that may result from single-objective optimization. For example, the algorithm can weigh whether to make more trips with medium-duty trucks to increase the loading rate or fewer trips with heavy-duty trucks to reduce fixed costs, thus providing a better decision-making basis. A multi-objective ant colony-genetic hybrid optimization algorithm is used to solve the problem, combining the global search capability of the ant colony algorithm and the local fine-grained search capability of the genetic algorithm, enabling more effective searching of non-dominated solution sets in a complex solution space. Ant colony optimization's iterative search helps explore a broad solution space, while the introduction of genetic algorithms further optimizes and maintains diversity in the solution set after a predetermined number of local searches. This prevents the algorithm from prematurely converging to a local optimum, increasing the likelihood of finding a high-quality allocation solution. It can handle large-scale, highly complex delivery scheduling problems and generate multiple allocation solutions with different trade-offs for users to choose from. The final allocation solution is selected based on user preferences. Decision-makers can make flexible choices based on actual operational needs (e.g., prioritizing cost control or ensuring high load rates during specific periods), making the delivery scheduling solution more adaptable and practical. This method, by comprehensively considering multiple vehicle types, time windows, dual-objective optimization, and hybrid intelligent algorithms, provides a more comprehensive, efficient, and flexible prefabricated component delivery scheduling optimization solution. It effectively improves resource utilization efficiency in multi-vehicle mixed delivery scenarios, reduces delivery costs, and enhances the practicality of the delivery solution.

[0028] The following is a detailed description of each step.

[0029] S1: Obtain parameter information of transport vehicles and construction site, wherein the parameter information of transport vehicles includes the type of transport vehicle, and the parameter information of construction site includes the delivery time window of construction site.

[0030] Specifically, the parameters of the transport vehicles include vehicle type (e.g., flatbed truck, van, heavy truck, etc.), maximum load capacity, unit transportation cost, fixed operating cost, vehicle size, and drivable area. The parameters of the construction sites include the site's geographical location, the types and quantities of prefabricated components required, and key delivery time windows. The prefabricated component manufacturing plant uses n vehicles of three types (large, medium, and small) to deliver the required components to m construction sites within a specified time. If the vehicle model selection is inappropriate, its fixed operating costs will increase, and there will be significant waste of vehicle loading space. If the vehicle organization is unreasonable, it can easily increase the extra time spent by the manufacturing plant waiting for loading and by the construction sites waiting for unloading. Furthermore, untimely delivery will lead to delays in construction progress, requiring the prefabricated component manufacturing plant to bear economic claims from the construction parties, thus incurring penalty costs. In reality, the prefabricated component manufacturing plant has multiple time windows limiting its delivery to multiple construction sites.

[0031] Before attempting to solve the aforementioned practical problems, certain assumptions are made: (1) The production plant has only one loading port, which can only load one vehicle at a time; (2) The construction site has only one unloading port, which can only unload one vehicle at a time; (3) The transportation phase is a point-to-point direct transportation between the factory and the construction site, and the vehicle returns directly to the production plant after unloading; (4) Except for the last delivery, the transport vehicles at each construction site must carry the maximum number of components allowed; (5) The loading time for the same vehicle model is the same, and the unloading time for the same vehicle model at the same construction site is the same; (6) The production plant has sufficient inventory of prefabricated components; (7) The transportation cost per unit kilometer is different for different vehicle models; (8) It is assumed that the vehicles travel at a constant speed, and traffic and other time consumption are not considered. Based on these assumptions, and to avoid the influence of non-essential factors, the optimization function is constructed and solved. When constructing the optimization function, the pre-obtained parameters and some intermediate parameter symbol definitions are shown in Table 1.

[0032] Table 1 Symbol Definitions

[0033]

[0034] S2: Based on the parameter information of the transport vehicles and the parameter information of the construction site, a bi-objective optimization function is constructed with the optimization objectives of minimizing delivery costs and maximizing the average vehicle loading rate. The delivery cost consists of the fixed usage cost of the vehicle, the transportation cost, and the penalty cost. The average vehicle loading rate is the sum of the single loading rates of all trips divided by the total number of trips.

[0035] Specifically, based on the distribution situation at the prefabricated component manufacturing plant, multi-objective optimization is adopted to improve efficiency, save costs, and more comprehensively meet actual needs. The optimization objectives are to minimize distribution costs and maximize the average vehicle loading rate. The dual-objective optimization function is as follows:

[0036] (1)

[0037] (2)

[0038] (3)

[0039] (4)

[0040] (5)

[0041] (6)

[0042] (7)

[0043] Equation (3) represents the amount given to the construction site. The arrival time of the kth delivery vehicle is later than the maximum time point of the time window. The delay time; Equation (4) represents the arrival time at the construction site. The time point of the kth vehicle; Equation (5) represents the arrival time at the construction site. The kth vehicle needs to wait because there are vehicles ahead of it unloading or vehicles in a queue to unload. As a determination coefficient, when the arrival time of the k-th vehicle is less than the departure time of the (k-1)-th vehicle, Equation (6) indicates that the kth vehicle leaves the construction site. The time is equal to the time the vehicle arrives at the construction site plus the waiting time for unloading plus the unloading time; Equation (7) represents the time given to the construction site. The time point at which the kth delivery vehicle returns to the production plant. Where F1 represents the delivery cost; F2 represents the average load factor. This represents the fixed cost per use of a transport vehicle of model r. This represents the unit transportation cost of the transport vehicle of model r; This represents the penalty cost coefficient for vehicles arriving late at the construction site; As a decision variable, when the construction site The kth car is of model r and numbered When the transport vehicle delivers, =1, otherwise It is 0; similarly, when the construction site The The vehicle is of model r and numbered When the transport vehicle delivers, =1, otherwise =0; This indicates the distance from the prefabricated component manufacturing plant to the construction site. The distance; Indicates to the construction site The arrival time of the kth delivery vehicle is later than the maximum time point of the time window. The delay time; m represents the total number of construction sites; n represents the total number of vehicles; R represents the set of vehicle types, such as {large, medium, small}; Indicates construction site The maximum number of the delivery train number; Indicates construction site The number of prefabricated components required; This indicates that vehicle model r transports prefabricated components to the construction site in one trip. Quantity; This indicates the maximum load capacity of the transport vehicle of model r. Indicates construction site The weight of a single prefabricated component required; INT() indicates rounding down; Indicates to the construction site The time when the kth delivery vehicle arrives at the construction site; Indicates the delay adjustment factor; Indicates to the construction site The time when the kth delivery vehicle departed from the production plant; This indicates that vehicles deliver components from the manufacturing plant to the construction site. The time required. Indicates to the construction site The time when the (k-1)th delivery vehicle leaves the construction site after unloading.

[0044] S3: A multi-objective ant colony-genetic hybrid optimization algorithm is used to solve the bi-objective optimization function. Under the conditions of meeting the delivery time window requirements of the construction site and preset constraints, multiple allocation schemes for various types of transport vehicles are obtained. The multi-objective ant colony-genetic hybrid optimization algorithm involves performing a preset number of local searches using the multi-objective ant colony optimization algorithm, followed by a single local search using the genetic algorithm. During algorithm execution, the ant colony algorithm simulates the behavior of ants searching for the shortest path, using pheromone accumulation to guide the search direction and explore different vehicle scheduling and path combinations. After a preset number of local searches using the ant colony algorithm, a single local search is performed using the genetic algorithm. The genetic algorithm optimizes the current solution set by simulating selection, crossover, and mutation operations in biological evolution, aiming to escape local optima and further improve the quality of the solution. The advantage of this hybrid algorithm is that it combines the advantages of both algorithms, enabling more effective searching of complex solution spaces and obtaining a series of non-dominated solutions, i.e., multiple allocation schemes.

[0045] Specifically, the constraint condition is

[0046] (8)

[0047] (9)

[0048] (10)

[0049] (11)

[0050] (12)

[0051] (13)

[0052] Equation (8) indicates that the arrival time of the delivery vehicle at the construction site is no earlier than the start time of construction at the construction site; Equation (9) indicates that the component factory needs to deliver components to each construction site, and the total number of delivery vehicles is less than the sum of the number of deliveries to each construction site; Equation (10) indicates that the total amount of precast components transported to a construction site must meet the demand of that construction site; Equation (11) indicates that the interval between two consecutive transport vehicles dispatched by the precast component production factory to a construction site is no less than the loading time of the previous transport vehicle; Equation (12) indicates that one delivery is completed by one vehicle; Equation (13) indicates that the total number of deliveries dispatched by the precast component production factory is equal to the total number of services provided to all construction sites. Among them, Indicates to the construction site The time when the kth delivery vehicle arrives at the construction site; Indicates construction site Start date of construction; Indicates construction site The maximum number of delivery vehicle numbers; m represents the total number of construction sites; n represents the total number of vehicles; R represents the set of vehicle types; As a decision variable, when the construction site The kth car is of model r and numbered The value is 1 when the delivery vehicle is present, and 0 otherwise. This indicates that vehicle model r transports prefabricated components to the construction site in one trip. Quantity; This indicates the maximum load capacity of the transport vehicle of model r. Indicates to the construction site The time when the kth delivery vehicle departed from the production plant; Indicates to the construction site The (k+1)th delivery vehicle's departure time from the production plant; Indicates to the construction site The loading time of the kth vehicle (model r) at the production plant.

[0053] The flowchart of the improved multi-objective ant colony-genetic hybrid optimization algorithm is shown in Figure 2. The parameters involved in the algorithm are initialized, and a series of control parameters are set to regulate the optimization process. For example, the population size parameter is set to 150 ants for parallel search, with a maximum iteration generation of 150 generations; the pheromone management parameter is set to a pheromone importance factor of 1.0, a volatility rate of 30%, and boundary value control; the heuristic weight parameter is set to a heuristic pheromone importance factor of 3.0, balancing multiple influencing factors; the search strategy parameter is set to a development-exploration balance parameter of 0.6, favoring the development mode; the local optimization parameter is set to a local search application probability of 50%, with a maximum number of attempts of 500; and the diversity maintenance parameter is set to an archive set capacity of 50, with a crowding calculation mechanism. The specific content of S3 is as follows.

[0054] S310: Construct an initial solution using a greedy strategy and a heuristic rule based on pheromone concentration.

[0055] Specifically, solve using a sequence of triples. It means that among them Let j be the construction site number, j be the vehicle number, and r be the vehicle type. Each ant constructs an initial solution through the following two stages.

[0056] S311: Employ a greedy strategy to allocate vehicles with large capacity to construction sites with high demand for prefabricated components, and update the remaining demand quantity at each site. The greedy strategy prioritizes vehicles that maximize loading rate or minimize cost when allocating vehicles.

[0057] S312: When the remaining demand is not met, vehicles are dynamically allocated to construction sites with remaining demand based on heuristic rules. When the remaining demand is not met, vehicles are dynamically allocated to construction sites with remaining demand based on heuristic rules (comprehensively considering pheromone concentration, parameters, transportation costs, loading rate, and time urgency). Following these two steps, a population of the corresponding size (i.e., the initial solution) is generated at the beginning of each iteration of the ant colony optimization algorithm. The pheromone parameters from the previous iteration are used in the next iteration.

[0058] Specifically, when the remaining demand is not met, the construction site with remaining demand is identified as the initial node. The transition probability is determined based on the pheromone concentration and the value of the bi-objective optimization function. Based on the magnitude of the transition probability, the next node is selected, thereby determining the dispatched vehicle number. The transition probability is...

[0059]

[0060]

[0061] in, This indicates that the kth vehicle of model r is used to supply the construction site. The probability of delivery being diverted. This indicates that the kth vehicle of model r is used to supply the construction site. The concentration of pheromones delivered This indicates that the kth vehicle of model r is used to supply the construction site. Heuristics about delivery Weighting factors representing pheromone concentration; Weighting factors representing heuristic information This represents the set of optional construction sites. This indicates that the s-th vehicle of model r is used to supply the construction site. The concentration of pheromones delivered This indicates that the s-th vehicle of model r is used to supply the construction site. Heuristics about delivery This indicates that the kth vehicle of model r is used to supply the construction site. Delivery costs This indicates that the kth vehicle of model r is used to supply the construction site. Expected load for delivery, This indicates the maximum load capacity of the transport vehicle of model r.

[0062] S320: According to the preset search strategy, the initial solution is updated and iterated by ant colony local search to obtain an optimized new solution.

[0063] Specifically, the preset search strategies include: swapping the site assignments of two delivery tasks, randomly inserting a new task, removing the task with the lowest load factor, and changing the vehicle type assignment of a task. If the local search is improved (i.e., delivery costs decrease or average load factor increases), the new solution is accepted. For example, the swap strategy: randomly select two delivery tasks, swap their site assignments, evaluate the improvement effect, and try 20 swap operations per generation. The insertion strategy: insert a new delivery task at a random position in the solution to try to increase delivery capacity; this strategy is enabled when the solution length is less than 18. The deletion strategy: identify the delivery task with the lowest load factor, remove it from the solution, and evaluate the impact on the overall objective. The vehicle type adjustment strategy: randomly select a delivery task and try using a different vehicle type, evaluating the changes in cost and load factor. Each strategy employs an acceptance criterion: a replacement is only accepted if the new solution is strictly superior to the original solution in terms of cost or load factor.

[0064] S330: Archive set update iteration: If the new optimized solution is not dominated by any solution in the archive set, then the new optimized solution is added to the archive set; if the new optimized solution is dominated by a solution in the archive set, then the new optimized solution is added to the archive set and the solution that dominates the new optimized solution in the archive set is removed; otherwise, the archive set remains unchanged.

[0065] In this context, assuming there are two solutions A and B, if both objective function values ​​of A are no worse than either of the two objective function values ​​of B, or at least one objective function value of A is significantly better than the corresponding objective function value of B, then A is said to dominate B. Furthermore, the archive set size can be limited to 50; if it exceeds this limit, the solution with the lowest crowding (worst diversity) is removed. For crowding, based on the distribution density in the objective space, the sum of the distances between each solution and its nearest neighbor is calculated; a larger distance indicates a greater contribution to diversity.

[0066] S340: Compare all solutions in the updated archive set with the global frontier solutions in the frontier solution set, identify non-dominant solutions, and update the frontier solution set iteratively.

[0067] Specifically, all solutions in the current generation are compared with the global front solution to identify solutions that are not dominated by each other, thus obtaining non-dominated solutions. Old solutions dominated by new solutions are removed, and new solutions that are not dominated are added, thereby updating the front solution set. The size of the front solution set is naturally controlled by the dominance relationship, without imposing hard limitations on the front solution set.

[0068] S350: If the number of ant colony local searches equals the preset number, then switch to a genetic algorithm for a genetic local search. Update the archive set and frontier solution set once, and reset the ant colony local search count to zero. The preset number of searches can be 5 as shown in Figure 2. After 5 local searches using the ant colony algorithm, execute a genetic algorithm for a local search. This step includes steps S351-S356.

[0069] S351: If the number of local searches by the ant colony is equal to the preset number, then based on the delivery cost of each solution in the frontier solution set generated in the current iteration, a roulette wheel blocking selection strategy is adopted to randomly select from the frontier solution set to determine the parent individual.

[0070] Specifically, a roulette wheel selection strategy is adopted. Parent individuals are randomly selected from the current non-dominated solution set based on the fitness of each Pareto front solution (front solution set), with lower costs resulting in higher fitness. Fitness is normalized and used as the selection probability to ensure that better solutions have a higher probability of being selected. Fitness transformation: The cost index is converted into selection probability; solutions with lower costs have a higher selection probability. Linear transformation ensures that all solutions have a chance to be selected. Probability normalization: The relative selection probability of each solution is calculated, ensuring that the sum of probabilities is 1, supporting random selection.

[0071] S352: For the selected parent individuals, generate the offspring solution using a single-point crossover method.

[0072] Specifically, crossover points are randomly selected in the solution sequence to ensure that the solutions after the crossover maintain a reasonable structure. The sequence before the first parent crossover point is combined with the sequence after the second parent crossover point to form a new child solution.

[0073] S353: Randomly change the allocation of construction sites, vehicles, or vehicle types in the sub-solutions with a preset probability to form new sub-solutions.

[0074] Specifically, the allocation of work sites, vehicles, or vehicle types for certain delivery tasks in the offspring solutions is randomly changed with a preset probability, thereby further increasing the diversity of solutions and preventing the algorithm from getting trapped in local optima. For example, the offspring solutions are randomly mutated with a 10% probability to increase population diversity. Work site mutation: Randomly changing the work site allocation for a delivery task. Vehicle mutation: Randomly changing the vehicle allocation for a delivery task. Vehicle type mutation: Randomly changing the vehicle type allocation for a delivery task.

[0075] S354: Repair the child solution, calculate the total amount of delivery allocated to each site, and for sites where the demand is not met, supplement the necessary delivery tasks to ensure that the demand of all sites is met, and obtain the repaired child solution.

[0076] Specifically, since crossover and mutation operations may result in sub-solutions failing to meet the needs of all work sites, a feasibility repair operation is performed on each sub-solution. The total allocated delivery volume for each work site is calculated, and for work sites with unmet needs, necessary delivery tasks are added to ensure that the needs of all work sites are met. This includes, for example, repairing, locally optimizing, and evaluating the generated sub-solutions. Demand Repair: Check the demand satisfaction of the mutated solution, add missing delivery tasks, and ensure that the needs of all work sites are met. Local Search: Perform local search optimization on the sub-solutions with a 50% probability. Quality Evaluation: Accurately calculate the total cost and average loading rate of the sub-solutions.

[0077] S355: Archive set update iteration: If the repaired child solution is not dominated by any solution in the archive set, then add the repaired child solution to the archive set; if the repaired child solution is dominated by a solution in the archive set, then add the repaired child solution to the archive set and remove the solution in the archive set that dominates the repaired child solution; otherwise, the archive set remains unchanged.

[0078] S356: Compare all solutions in the updated archive set with the global frontier solutions in the frontier solution set, identify solutions that do not dominate each other, update the frontier solution set, and reset the number of local searches of the ant colony to zero and start accumulating again.

[0079] Specifically, all solutions in the current generation are merged, and the frontier solution set is updated. Non-dominated solution filtering: All solutions in the current generation are compared with the global frontier solution to identify solutions that are not mutually dominant. Frontier set update: Old solutions dominated by new solutions are removed, and new solutions that are not dominated are added.

[0080] By performing a predetermined number of local searches using the ant colony algorithm, followed by a final local search using a genetic algorithm, the advantages of both algorithms are effectively combined. The ant colony algorithm excels at exploring the search space, while the genetic algorithm excels at developing and optimizing within discovered regions. When the ant colony algorithm might get stuck in local optima, the crossover and mutation operations of the genetic algorithm can escape these local optima and introduce new solutions. The roulette wheel selection strategy prioritizes the use of parent individuals with higher fitness for reproduction, thus accelerating convergence. Single-point crossover and random mutation operations ensure the diversity and exploratory nature of new solutions. Furthermore, the repair mechanism for offspring solutions ensures that newly generated solutions always meet actual delivery requirements, avoiding the generation of invalid solutions. This hybrid optimization strategy allows the algorithm to maintain global search capabilities while also possessing strong local optimization capabilities, enabling it to find high-quality prefabricated component delivery scheduling schemes more efficiently.

[0081] S360: If the frontier solution set changes, the pheromone concentration is updated according to the preset pheromone evaporation rate to obtain the updated pheromone concentration, which is used to solve the bi-objective optimization function in the next round.

[0082] Specifically, upper and lower limits for pheromone are added during the evaporation process to prevent excessive decay or explosion of pheromone concentration. If a better solution is found through local search, the pheromone increment is multiplied by a reward coefficient. For example, the pheromone concentration is limited to between 0.01 and 10.0 to avoid extreme values ​​affecting the search balance. Pheromones are only enhanced on delivery paths corresponding to non-dominated solutions. The enhancement strength is positively correlated with the quality of the solution; solutions with lower costs and higher loading rates provide stronger pheromone enhancements. All pheromone paths evaporate at a rate of 30% to prevent premature convergence of the algorithm.

[0083] The updated pheromone concentration is

[0084]

[0085] in, This represents the pheromone concentration at iteration number t+1; This represents the pheromone concentration at iteration number t. For pheromone evaporation rate; Indicates the minimum concentration of pheromones. This indicates the maximum pheromone concentration; clip() is a numerical processing function that constrains the value within a specified range of minimum and maximum values.

[0086] Furthermore, algorithm parameters can be dynamically adjusted based on the search status. If the number of non-dominated solutions at the forefront is small, the exploration probability is reduced and the pheromone evaporation rate is increased; if the number of non-dominated solutions at the forefront is large, the exploration probability is increased and the pheromone evaporation rate is decreased, thus strengthening the local search. For example, when there are fewer than 5 non-dominated solutions and more than 20 iterations, exploration behavior is increased (development parameters are reduced, and the evaporation rate is increased); when there are more than 10 non-dominated solutions, development behavior is increased (development parameters are increased, and the evaporation rate is decreased). By adjusting the parameters, a balance between development and exploration is maintained, avoiding premature convergence or random walks.

[0087] S370: Until the number of iterations exceeds the preset total number of iterations or no new solutions are added to the archive set after multiple consecutive iterations, the global frontier solutions in the frontier solution set are used as multiple allocation schemes for various types of transport vehicles.

[0088] Specifically, as shown in Figure 2, when the number of iterations GEN is less than a set value (i.e., the preset total number of iterations), the iteration ends, the frontier solution set and the archive set are merged, and solutions that do not dominate each other are selected from the merged set to ensure the quality of the final solution set. The optimal final solution set is then output. During output, the final solution set can also be sorted, for example, by ascending order of total cost.

[0089] By combining a greedy strategy with heuristic rules based on pheromone concentration to construct the initial solution, the algorithm can start its search from a relatively optimal starting point, effectively avoiding blind exploration and thus accelerating the convergence speed. Secondly, the dynamic update and iteration mechanism of the archive set and the frontier solution set ensures the effective maintenance and management of a high-quality non-dominated solution set during multi-objective optimization, preventing the loss of excellent solutions and promptly reflecting the optimal Pareto front reached by the algorithm. More importantly, the alternation between ant colony local search and genetic algorithm local search fully utilizes the global exploration capability of the ant colony algorithm and the local development capability of the genetic algorithm, enabling the algorithm to effectively escape local optima and find better and more diverse distribution schemes in a broader solution space. Furthermore, the dynamic pheromone update mechanism based on changes in the frontier solution set effectively guides the search direction of the ant colony algorithm, making it more inclined to explore promising regions, further improving search efficiency and solution quality. These improvements work together to enable the method of this application to generate a more efficient and accurate prefabricated component delivery and allocation plan that takes into account both delivery costs and average vehicle loading rates, while meeting the delivery time window requirements and preset constraints of construction sites, thus providing users with better decision support.

[0090] S4: Select the allocation plan based on user preferences and deliver the prefabricated components.

[0091] Specifically, for example, after obtaining multiple non-dominated solutions (e.g., one solution has the lowest cost but a slightly lower loading rate, and another solution has the highest loading rate but a slightly higher cost), the user can select the most suitable solution based on current business priorities (e.g., whether to prioritize cost control or resource utilization). Once a solution is selected, the loading, transportation, and delivery of prefabricated components can be carried out according to that solution.

[0092] The above method is applied to a specific example, and the application process and analysis results are as follows.

[0093] A prefabricated component manufacturing plant has 3 small transport vehicles, 2 medium-sized transport vehicles, and 1 large transport vehicle. Parameters such as the maximum load capacity, transportation cost, and fixed operating cost of the vehicles are shown in Table 2. This manufacturing plant needs to deliver prefabricated components to 4 surrounding construction sites. The delivery distance and round-trip time between the manufacturing plant and the construction sites are shown in Table 3. Parameters such as the number of components required by each construction site and the weight of each component are shown in Table 4. Parameters such as the unloading time and demand time window for different vehicle types at different construction sites are shown in Table 5. Unit penalty cost coefficient. =1.5 yuan / minute, and the factory starts delivering at 8:00 AM.

[0094] The precast component production plant delivers to the construction site sequentially according to the time window, and the optimal delivery plan with 20 planned trips is obtained. Based on the proposed dual-objective optimization function, the delivery cost is calculated to be 9076.20 yuan, and the average loading rate is 92.63%.

[0095] Table 2 Vehicle-related data

[0096]

[0097] Table 3. Construction Site Related Data

[0098]

[0099] Table 4 Delivery-related data

[0100]

[0101] Table 5 Unloading Time

[0102]

[0103] Based on the distribution data from the precast component manufacturing plant, an improved multi-objective ant colony-genetic hybrid optimization algorithm was used to solve the optimization objective function. Eleven unique Pareto optimal solutions (i.e., multiple allocation schemes) were obtained, and the spatial distribution of the solutions is shown in Figure 3. The objective function values ​​of each Pareto optimal solution are shown in Table 6.

[0104] Table 6. Values ​​of each objective function

[0105]

[0106] According to the information in Table 6, all optimal solutions have lower delivery costs than the original delivery plan and higher average loading rates. All solutions in the Pareto solution set can be selected as feasible options; the choice depends on the decision-maker's preference. Here, prioritizing cost savings, the delivery plan with a delivery cost of 8677.50 yuan and an average loading rate of 93.79% is selected for demonstration. The dispatch sequence and schedule for the production plant are shown in Table 7, and the dispatch sequence and schedule for each construction site are shown in Table 8. In the tables, vehicle numbers 1, 2, and 3 are small transport vehicles, 4 and 5 are medium-sized transport vehicles, and 6 is a large transport vehicle.

[0107] Table 7 Production Plant Outbound Sequence and Schedule

[0108]

[0109] Table 8. Order and Schedule of Vehicle Departure from Each Construction Site

[0110]

[0111] As can be seen from the dispatch sequence and timing of the above scheme, under the conditions of time window constraints and multiple vehicle types, all deliveries to construction sites are completed within the available pickup time at each site, ensuring the effectiveness of the delivery. Compared to the original delivery scheme requiring 20 dispatches, this scheme only used 18 dispatches to meet the delivery needs, proving the feasibility and practicality of the proposed dual-objective optimization function, which is beneficial to improving the industry competitiveness of precast component manufacturing plants.

[0112] Based on the delivery data from the precast component manufacturing plant, the original multi-objective ant colony optimization algorithm was used to solve the optimization objective function, and only three non-repeating Pareto optimal solutions were obtained. The spatial distribution of the solutions is shown in Figure 4, and the objective function values ​​of each Pareto optimal solution are shown in Table 9.

[0113] Table 9. Values ​​of each objective function

[0114]

[0115] Therefore, it is evident that the original multi-objective ant colony optimization algorithm is prone to getting trapped in local optima during the solution process, resulting in insufficient solution diversity. Furthermore, its performance in terms of objective function values ​​is inferior to the improved multi-objective ant colony-genetic hybrid optimization algorithm. The improved algorithm, by enhancing global search capabilities and enriching the solution set structure, effectively overcomes the problem of premature convergence in the original algorithm, significantly improving solution diversity and optimization performance.

[0116] As shown in Figure 5, this application provides a prefabricated component delivery scheduling optimization system based on time windows and multiple vehicle models, used to implement the above-mentioned prefabricated component delivery scheduling optimization method based on time windows and multiple vehicle models. The system includes:

[0117] The data acquisition module 10 is used to acquire parameter information of transport vehicles and parameter information of construction sites. The parameter information of transport vehicles includes the type of transport vehicle, and the parameter information of construction sites includes the delivery time window of the construction site.

[0118] The optimization function construction module 20 is used to construct a bi-objective optimization function based on the parameter information of the transport vehicles and the parameter information of the construction site, with the optimization objectives of minimizing the delivery cost and maximizing the average vehicle loading rate. The delivery cost consists of the fixed usage cost of the vehicle, the transportation cost and the penalty cost, and the average vehicle loading rate is the sum of the single loading rates of all trips divided by the total number of trips.

[0119] The optimization function solution module 30 is used to solve the bi-objective optimization function using a multi-objective ant colony-genetic hybrid optimization algorithm. Under the conditions of meeting the delivery time window requirements of the construction site and preset constraints, it obtains multiple dispatch schemes for various types of transport vehicles. The multi-objective ant colony-genetic hybrid optimization algorithm is transformed into a genetic algorithm for a local search after each preset number of local searches using the multi-objective ant colony optimization algorithm.

[0120] The allocation scheme determination module 40 is used to select an allocation scheme based on user preferences and to deliver prefabricated components.

[0121] In this embodiment, the beneficial effects of the prefabricated component delivery scheduling optimization system based on time windows and multiple vehicle models are similar to those of the prefabricated component delivery scheduling optimization method based on time windows and multiple vehicle models described above, and will not be repeated here.

[0122] An electronic device provided in this application includes a memory and a processor; the memory is used to store a computer program; the processor is used to implement the above-described optimization method for prefabricated component delivery scheduling based on time windows and multiple vehicle models when executing the computer program.

[0123] This application provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the prefabricated component delivery scheduling optimization method based on time windows and multiple vehicle models as described above.

[0124] In this embodiment, the beneficial effects of the electronic device and the computer-readable storage medium are similar to those of the above-described optimization method for prefabricated component delivery scheduling based on time windows and multiple vehicle models, and will not be repeated here.

[0125] The present invention describes electronic devices that can serve as servers or clients of this application, which are examples of hardware devices that can be applied to various aspects of this application. Electronic devices are intended to represent various forms of digital electronic computer devices, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic devices can also represent various forms of mobile devices, such as personal digital assistant devices, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the application described and / or claimed herein.

[0126] Electronic devices include a computing unit that can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) or loaded from a storage unit into random access memory (RAM). The RAM can also store various programs and data required for device operation. The computing unit, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0127] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc. In this application, the separately described modules may or may not be physically separate. Some or all of the modules can be selected to achieve the purpose of the embodiments of this application according to actual needs. Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing unit, or each module can exist physically separately, or two or more modules can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0128] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0129] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for optimizing the delivery scheduling of prefabricated components based on time windows and multiple vehicle models, characterized in that, include: The process involves acquiring parameter information for transport vehicles and construction sites. The transport vehicle parameters include vehicle type, and the construction site parameters include delivery time windows. Based on these parameters, a bi-objective optimization function is constructed, aiming to minimize delivery costs and maximize average vehicle loading rate. Delivery costs consist of fixed vehicle usage costs, transportation costs, and penalty costs. The average vehicle loading rate is the sum of single-trip loading rates for all trips divided by the total number of trips. A multi-objective ant colony-genetic hybrid optimization algorithm is used to solve the bi-objective optimization function. Under the conditions of meeting the construction site delivery time window requirements and preset constraints, multiple allocation schemes for various types of transport vehicles are obtained. The multi-objective ant colony-genetic hybrid optimization algorithm involves performing a preset number of local searches using the multi-objective ant colony optimization algorithm, followed by a single local search using the genetic algorithm. Based on user preferences, an allocation scheme is selected, and prefabricated components are delivered. The bi-objective optimization function is... Where F1 represents delivery cost; F2 represents average load factor; This represents the fixed cost per use of a transport vehicle of model r. This represents the unit transportation cost of the transport vehicle of model r; This represents the penalty cost coefficient for vehicles arriving late at the construction site; As a decision variable, when the construction site The kth car is of model r and numbered When the transport vehicle delivers, =1, otherwise It is 0; similarly, when the construction site The The vehicle is of model r and numbered When the transport vehicle delivers, =1, otherwise =0; This indicates the distance from the prefabricated component manufacturing plant to the construction site. The distance; Indicates to the construction site The arrival time of the kth delivery vehicle is later than the maximum time point of the time window. The delay time; m represents the total number of construction sites; n represents the total number of vehicles; R represents the set of vehicle types; Indicates construction site The maximum number of the delivery train number; Indicates construction site The number of prefabricated components required; This indicates that vehicle model r is used to transport prefabricated components to the construction site in one trip. Quantity; This indicates the maximum load capacity of the transport vehicle of model r. Indicates construction site The weight of a single prefabricated component required; INT() indicates rounding down; Indicates to the construction site The time when the kth delivery vehicle arrives at the construction site; Indicates the delay adjustment factor; Indicates to the construction site The time when the kth delivery vehicle departed from the production plant; This indicates that vehicles deliver components from the manufacturing plant to the construction site. The time required.

2. The prefabricated component delivery scheduling optimization method based on time windows and multiple vehicle models as described in claim 1, characterized in that, The multi-objective ant colony-genetic hybrid optimization algorithm is used to solve the bi-objective optimization function. Under the conditions of meeting the delivery time window requirements of the construction site and preset constraints, multiple allocation schemes for various types of transport vehicles are obtained, including: constructing an initial solution using a greedy strategy and a heuristic rule based on pheromone concentration; updating and iterating the initial solution using ant colony local search according to a preset search strategy to obtain an optimized new solution; updating and iterating the archive set: if the optimized new solution is not dominated by any solution in the archive set, then the optimized new solution is added to the archive set; if the optimized new solution is dominated by a solution in the archive set, then the optimized new solution is added to the archive set, and the solution that dominates the optimized new solution in the archive set is removed; otherwise, the archive set remains unchanged; and updating the current generation... All solutions in the newly archived set are compared with the global frontier solutions in the frontier solution set to identify non-dominant solutions and update the frontier solution set. If the number of ant colony local searches equals the preset number, the algorithm is transformed into a genetic local search, updating the archived set and the frontier solution set once, and the number of ant colony local searches is reset to zero and re-accumulated. If the frontier solution set changes, the pheromone concentration is updated according to the preset pheromone evaporation rate to obtain the updated pheromone concentration, which is used to solve the bi-objective optimization function in the next round. This continues until the number of iterations exceeds the preset total number or no new solutions are added to the archived set after multiple consecutive iterations. Then, the global frontier solutions in the frontier solution set are used as multiple allocation schemes for various types of transport vehicles.

3. The prefabricated component delivery scheduling optimization method based on time windows and multiple vehicle models as described in claim 2, characterized in that, The method of constructing an initial solution using a greedy strategy and a heuristic rule based on pheromone concentration includes: using a greedy strategy to allocate vehicles with large capacity to construction sites with large demand for prefabricated components and updating the remaining demand quantity of the construction sites; and dynamically scheduling vehicles to construction sites with remaining demand when the remaining demand quantity is not met, based on heuristic rules.

4. The prefabricated component delivery scheduling optimization method based on time windows and multiple vehicle models as described in claim 3, characterized in that, The method of dynamically scheduling vehicles to construction sites with remaining demand when the remaining demand is not met, based on heuristic rules, includes: determining the construction site with remaining demand as the initial node when the remaining demand is not met; determining the transition probability based on the pheromone concentration and the value of the bi-objective optimization function; and selecting the next node based on the magnitude of the transition probability, thereby determining the dispatched vehicle number.

5. The prefabricated component delivery scheduling optimization method based on time windows and multiple vehicle models as described in claim 4, characterized in that, The transition probability is in, This indicates that the kth vehicle of model r is used to supply the construction site. The probability of delivery being diverted. This indicates that the kth vehicle of model r is used to supply the construction site. The concentration of pheromones delivered This indicates that the kth vehicle of model r is used to supply the construction site. Heuristics about delivery Weighting factors representing pheromone concentration; Weighting factors representing heuristic information This represents the set of optional construction sites. This indicates that the s-th vehicle of model r is used to supply the construction site. The concentration of pheromones delivered This indicates that the s-th vehicle of model r is used to supply the construction site. Heuristics about delivery This indicates that the kth vehicle of model r is used to supply the construction site. Delivery costs This indicates that the kth vehicle of model r is used to supply the construction site. Expected load for delivery, This indicates the maximum load capacity of the transport vehicle of model r.

6. The prefabricated component delivery scheduling optimization method based on time windows and multiple vehicle models as described in claim 2, characterized in that, The preset search strategies include: swapping the site assignments of two locations, randomly inserting a new task, deleting the task with the lowest loading rate, and changing the vehicle type assignment of a task.

7. The prefabricated component delivery scheduling optimization method based on time windows and multiple vehicle models as described in claim 2, characterized in that, If the number of local searches by the ant colony equals the preset number, then a genetic algorithm is used for genetic local search. The archive set and the frontier solution set are updated and iterated once, and the number of local searches by the ant colony is reset to zero and re-accumulated. This includes: if the number of local searches by the ant colony equals the preset number, then based on the delivery cost of each solution in the frontier solution set generated in the current iteration, a roulette wheel selection strategy is used to randomly select a parent individual from the frontier solution set; for the selected parent individual, a single-point crossover method is used to generate offspring solutions; the allocation of construction sites, vehicles, or vehicle types in the offspring solutions is randomly changed with a preset probability to form new offspring solutions; the offspring solutions are repaired, and the delivery allocated to each construction site is statistically analyzed. For the total quantity, for construction sites with unmet needs, supplement necessary delivery tasks to ensure that the needs of all construction sites are met, and obtain the repaired child solution; Archive set update iteration: If the repaired child solution is not dominated by any solution in the archive set, then add the repaired child solution to the archive set; if the repaired child solution is dominated by a solution in the archive set, then add the repaired child solution to the archive set and remove the solution in the archive set that dominates the repaired child solution; otherwise, the archive set remains unchanged; compare all solutions in the current generation's updated archive set with the global front solution in the front solution set, identify solutions that do not dominate each other, update the front solution set, and reset the number of ant colony local searches to zero and start accumulating again.

8. The prefabricated component delivery scheduling optimization method based on time windows and multiple vehicle models as described in claim 2, characterized in that, The updated pheromone concentration is in, This represents the pheromone concentration at iteration number t+1; This represents the pheromone concentration at iteration number t. For pheromone evaporation rate; Indicates the minimum concentration of pheromones. This indicates the maximum pheromone concentration; clip() is a numerical processing function that constrains the value within a specified range of minimum and maximum values.

9. The prefabricated component delivery scheduling optimization method based on time windows and multiple vehicle models as described in claim 1, characterized in that, The constraint condition is: in, Indicates construction site Start date of construction; Indicates to the construction site The loading time of the kth vehicle (model r) at the production plant.

Images