Solid waste plastic production task scheduling and resource balanced distribution method

Abstract

The invention belongs to the technical field of information, and particularly relates to a solid waste plastic production task scheduling and resource balanced distribution method which is applied to waste plastic regeneration enterprises. The method comprises the steps of firstly collecting order, resource and production process data and constructing a multi-objective optimization model, and then improving a differential evolution algorithm from three aspects of adaptive parameter adjustment, an elitism strategy and multi-population coevolution to solve the model. The generated optimization scheme can dynamically adjust production through implementation and real-time monitoring. The method can effectively solve the contradiction between heavy production tasks and limited resources of enterprises, improves the production efficiency and the resource utilization rate, and has high application value.

Description

Technical Field

[0001] The present invention belongs to the field of production management of waste plastic recycling enterprises, and specifically provides a method for scheduling solid waste plastic production tasks and balancing resource allocation. Background Art

[0002] In the waste plastic recycling industry, as market demand grows, companies often need to take on multiple orders simultaneously to meet market supply. However, each order has specific delivery times and product quality requirements, making production tasks complex and diverse. Furthermore, companies have limited resources, such as production equipment and manpower. Irrational production task scheduling and resource allocation can lead to a series of problems. For example, due to excessive task concentration, some equipment may operate at high load for extended periods, shortening equipment lifespan and increasing maintenance costs. This can also lead to production bottlenecks, impacting overall production progress. Meanwhile, other equipment may remain idle, resulting in a waste of resources. Furthermore, failure to deliver orders on time can damage a company's reputation and reduce its market competitiveness.

[0003] Traditional production task scheduling and resource allocation methods rely primarily on manual experience or simple rules, making them incapable of handling complex and volatile order and resource situations. Differential evolution, as a highly efficient evolutionary algorithm, offers advantages in solving optimization problems. However, when dealing with complex problems such as production task scheduling and balanced resource allocation, standard differential evolution algorithms are prone to becoming stuck in local optima and exhibit slow convergence. Therefore, improvements to differential evolution algorithms are necessary to better meet the production management needs of waste plastic recycling companies and achieve optimal scheduling of production tasks and balanced resource allocation. Summary of the Invention

[0004] The present invention provides a waste plastic recycling production sequence optimization method based on a tabu search algorithm, which aims to resolve the contradiction between heavy production tasks and limited resources in enterprises that simultaneously undertake multiple waste plastic recycling orders. By rationally scheduling production tasks and evenly allocating resources, the overall production efficiency and resource utilization rate of the enterprise are improved, including:

[0005] Data collection: Collect data on production operations, energy consumption, quality impact, equipment maintenance, production environment and market demand;

[0006] Problem modeling: Under the premise of meeting quality, link relationships, equipment maintenance and market demand, with operation selection, time adjustment and market response strategy as decision variables, a model is constructed to find the optimal production sequence. The goal is to minimize energy consumption, maximize efficiency and respond to the market. Let the production sequence solution vector be S = [s1, s2, ..., s n], the quality of the solution is evaluated by the objective function Z = w1E(S) + w2P(S) + w3M(S), where E(S) is the energy consumption function, P(S) is the production efficiency function, M(S) is the market response function, and w1, w2, and w3 are weight coefficients determined according to the company's production goals and market conditions;

[0007] Algorithm implementation: Initialize the solution and taboo table, select the optimal solution from the neighborhood during iteration, and generate neighborhood solutions by changing the operation type, sequence, number of operations, operation conditions, or market strategy; remove taboos based on conditions, and dynamically maintain the taboo table based on solution quality, market stability, etc.

[0008] Solution implementation: Dynamically determine the number of iterations based on factors such as production scale to obtain the optimal sequence, combine production management with market forecasting systems to arrange production, monitor in real time and adjust according to market conditions.

[0009] Furthermore, the taboo table has a dynamic deadline based on solution quality, market and environmental changes, and whether it will be removed will be determined based on re-evaluation standards when it expires.

[0010] Furthermore, the generation of neighborhood solutions takes into account the impact on the previous and subsequent links, equipment, environment and market, including operation type changes, sequence exchanges, number fine-tuning and market strategy adjustments.

[0011] Furthermore, the present invention also provides a method for optimizing the allocation of waste plastic recycling resources based on the whale optimization algorithm, comprising:

[0012] Data collection and problem definition: Collect data related to resource demand, cost, loss, suppliers and supply chain, and define the resource allocation problem that seeks the best efficiency, energy consumption, cost and supply chain collaboration under resource and supply chain constraints.

[0013] Algorithm application: The whale position represents the resource allocation plan, which is adjusted according to the encirclement, bubble net attack and prey search behavior. In the spiral update position of the bubble net attack phase, the whale position update formula is X i (t+1)=X*(t)+D·e bl ·cos(2πl), where X i (t+1) is the position of the i-th whale at time t+1, X*(t) is the current optimal solution position, D = |C·X * (t)-X i (t)|, C is a random vector, b is a constant controlling the spiral shape, and l is a random number between (-1, 1). Adjustment parameters at each stage dynamically change based on factors such as resources, production, costs, and supply chain. The search range and step size are adaptively adjusted based on population, market, and supply chain risks.

[0014] Solution implementation: Determine the optimal solution with the number of iterations based on factors such as resources and supply chain, allocate resources using intelligent resource management and supply chain collaboration systems, and monitor and adjust in real time based on supply chain changes.

[0015] Furthermore, in the encirclement phase, whales move towards the optimal solution according to the proportions related to resources, suppliers, and supply chains. The proportions are dynamically adjusted based on iterations, resources, costs, link priorities, and supply chain risks.

[0016] Furthermore, during the bubble network attack phase, the contraction amplitude and spiral parameters change dynamically depending on resources, production, cost, supply chain, and iteration coefficient.

[0017] Furthermore, the present invention also provides a method for optimizing the formulation of waste plastic recycled products based on a differential evolution algorithm, comprising:

[0018] Data collection and model building: Collect data on raw materials, product performance, cost, energy consumption, market and competition, and build multiple models to describe the relationship between formula and various factors.

[0019] Algorithm application: Taking the raw material ratio as the decision variable, the formula is optimized through mutation, crossover, and selection operations. In the mutation operation, the formula for generating mutant individuals is V i，G =X r1，G +F×(X r2，G -X r3，G ), where V i，G is the i-th mutant individual of the G-th generation, X r1，G 、X r2，G 、X r3，G are three different individuals randomly selected from the Gth generation population, and F is the variation factor. Each operation is adaptively adjusted based on factors such as raw materials, performance, market, and competition. The fitness function integrates factors such as product quality, cost, energy consumption, market response, and competitive advantage.

[0020] Program implementation: Determine the optimal formula through the number of iterations based on factors such as product, market, and competition, utilize multiple monitoring and intelligence systems to produce and fine-tune based on actual conditions, market, competition, and environmental policies.

[0021] Furthermore, the mutation factor is adjusted dynamically based on population, raw materials, performance, market, and competition, enhancing global search in the early stages of evolution and local search in the later stages.

[0022] Furthermore, crossover factors are dynamically adjusted based on individual fitness, performance, market, competition, and environmental requirements to determine whether genes are retained or updated.

[0023] Furthermore, the fitness function integrates product quality, market demand, competitive advantage, production cost, energy consumption, market penalty and environmental cost, and the weight and calculation method of each score are determined according to the corresponding factors and models.

[0024] Beneficial effects

[0025] Improve order delivery reliability: Through precise production task scheduling, fully considering factors such as order delivery time and product requirements, this method effectively ensures that orders are delivered on time. Optimize resource utilization efficiency: Achieve balanced allocation of various resources, reduce equipment idleness and overuse. Enhance production flexibility and adaptability: In the face of changes in production, such as order changes and equipment failures, the adaptive parameter adjustment mechanism enables the algorithm to respond quickly and re-plan production tasks and resource allocation. Without complex resetting, a new optimal solution can be found to ensure stable and efficient production operation and reduce the impact of emergencies on production. Improve production efficiency: Improve the algorithm's global search and rapid convergence capabilities to help companies find better production solutions and reduce waste and delays in the production process. In links such as injection molding, companies have more reasonable production arrangements, significantly improved overall production efficiency, and created more economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 Data collection and problem modeling flow chart

[0027] Figure 2 Improved differential evolution algorithm flow chart

[0028] Figure 3 Optimization plan implementation and monitoring flow chart DETAILED DESCRIPTION

[0029] 1. Data collection and problem modeling implementation

[0030] Data collection implementation

[0031] Order Data Collection: A dedicated order management system is designed. Upon receiving a new order, sales personnel enter detailed order information, including order number, product type, quantity, delivery time, and product quality requirements. The system automatically checks the format and completeness of the entered data to ensure accuracy. Simultaneously, the system maintains communication with customers, promptly obtains order change information, and updates order data.

[0032] Resource Data Collection: For production equipment, establish an equipment archive to record basic equipment information (type, model, production capacity, etc.), maintenance records (maintenance time, maintenance content, maintenance personnel, etc.), and operating status (collected in real time through equipment sensors or manually entered periodically). Utilize the human resources management system to compile employee information, including personal information, skill levels, and work schedules. Energy resource data, including energy supply contract information and real-time energy consumption data, is obtained through data integration with energy suppliers and internal energy metering equipment.

[0033] Production process data collection: Organize production technicians to conduct a detailed review of each process step in the waste plastic recycling production process, recording the operating procedures, process parameter requirements (such as temperature, pressure, time, etc.), and resource consumption (equipment usage time, raw material usage, etc.) for each process step. Establish a process database for each product type to accurately obtain the corresponding process information during production task scheduling and resource allocation.

[0034] Problem Modeling Implementation

[0035] Multi-objective optimization model construction: Based on the goals of on-time order delivery, balanced resource utilization, and minimum production cost, corresponding objective functions are defined. For on-time order delivery, overdue orders are handled accordingly to reflect their adverse impact on optimization. For balanced resource utilization, the dispersion of various indicators (such as equipment utilization rate and employee working hours) is calculated to assess the balance of resource utilization. For minimum production cost, the cost of each resource is comprehensively calculated. These objective functions are combined into a fitness function through weighted summation. The weight coefficients are adjusted according to the company's actual situation to comprehensively evaluate different production task scheduling and resource allocation schemes.

[0036] Solution vector representation: The production task for each order is broken down according to the production process, with each subtask corresponding to a decision variable. For example, for an order that includes an injection molding process, the decision variables can be represented as the injection molding equipment number, the operator number, the injection molding start time, the injection molding process parameters, etc. The subtask decision variables of all orders form a solution vector, which represents a production task scheduling and resource allocation plan. To ensure the legitimacy of the solution vector, a series of constraints are formulated, such as equipment production capacity constraints (to ensure that the production tasks assigned to a certain equipment do not exceed its production capacity), personnel skill constraints (to ensure that the operators have the corresponding skills), and time constraints (to ensure that the time schedule of production tasks is reasonable and does not conflict).

[0037] (2) Improved implementation of differential evolution algorithm

[0038] Initialize population implementation

[0039] Solution vector generation: Based on the solution space scope and constraints, a random number generation algorithm is used to generate an initial solution vector. For example, for the equipment allocation decision variable, a random number is selected within the available equipment number range; for the time decision variable, a reasonable start and end time is randomly generated within the order delivery time limit. After the solution vector is generated, it is checked for validity according to the constraints. If the solution vector does not meet the constraints, a repair algorithm is used to correct it. For example, if a task assigned to a device exceeds its production capacity, the task is reassigned to another device.

[0040] Population size determination: Determine the appropriate population size based on the complexity of the problem and the computing resources. A population size that is too small may limit the algorithm's search space and easily lead to local optima. A population size that is too large may increase computing time and resource consumption. Determine the appropriate population size through experimentation and experience.

[0041] Subpopulation division (if multi-population coevolution is used): If a multi-population coevolution strategy is used, the population is divided into multiple subpopulations. The division of subpopulations can be done randomly or based on certain characteristics (such as individual fitness values, the distribution of decision variables, etc.).

[0042] Fitness calculation implementation

[0043] Objective Function Calculation: Based on the constructed multi-objective optimization model, the on-time order delivery, balanced resource utilization, and production cost performance of each solution vector are calculated. On-time order delivery is evaluated based on the actual order completion time and delivery time; balanced resource utilization is evaluated based on the distribution of resource usage; and production cost is calculated by statistically analyzing the consumption of various resources. These evaluation results are combined to calculate the fitness value of each solution vector.

[0044] Comprehensive fitness assessment: Based on the importance of different objectives, the calculation results of each objective function are weighted and summed to obtain the final fitness value. Based on the company's key focus, the weights of each objective are reasonably adjusted to make the fitness value more consistent with the company's optimization needs.

[0045] Mutation operation implementation

[0046] Mutation process: For each individual (solution vector) in the population, three different individuals are randomly selected from the population according to the adaptively adjusted mutation factor, and a mutant individual is generated through a difference operation. This mutation operation provides new possibilities for searching for a better solution.

[0047] Impact of mutation: Mutation operations can cause the solution vector to change within a certain range, helping the algorithm to escape from local optimality and explore a wider solution space, creating conditions for finding better production task scheduling and resource allocation solutions.

[0048] Cross-operation implementation

[0049] Crossover process: The mutant individual is cross-pollinated with the original individual according to the adaptively adjusted crossover factor to generate a test individual. Based on the set crossover rules, the element source of the test individual is determined, so that the test individual combines partial information of the mutant individual and the original individual.

[0050] The significance of crossover: The crossover operation can combine the excellent characteristics of different individuals, increase the diversity of the population, and improve the probability of finding a better solution.

[0051] Select Action Implementation

[0052] Selection process: Compare the fitness values ​​of the test individuals with those of the original individuals and select the individuals with better fitness to enter the next generation of the population. This ensures that the quality of individuals in the population continues to improve and evolve towards the optimal solution.

[0053] Elite retention: At the same time, several individuals with the best fitness in the current population are directly copied to the next generation population, retaining the currently found excellent solutions and accelerating the convergence of the algorithm.

[0054] Multi-population co-evolution implementation

[0055] Information exchange process: If a multi-population co-evolution strategy is adopted, information exchange between sub-populations is carried out regularly. Some of the best individuals are selected from each sub-population and transferred to other sub-populations to replace some of the poorer individuals in other sub-populations, promoting co-evolution between sub-populations and avoiding falling into local optimal solutions.

[0056] Collaborative effect: By sharing information between sub-populations, search information from different regions can be integrated, improving the algorithm's global search capability and helping to find better production task scheduling and resource allocation solutions.

[0057] Equipment layout optimization based on cuckoo search algorithm

[0058] Company Background

[0059] Large-scale waste plastic recycling companies typically have large and complex production workshops equipped with numerous different types of equipment, such as crushers, washers, dryers, and granulators. These devices closely interact with each other through material flow and production collaboration. However, traditional equipment layouts often lack scientific planning, resulting in circuitous material transportation routes within the workshop, increasing transportation time and energy consumption while also reducing the efficiency of equipment collaboration. Therefore, optimizing equipment layout is crucial for improving production efficiency and reducing energy consumption.

[0060] Specific implementation

[0061] Data Collection and Problem Modeling: Measure the workshop's spatial dimensions in detail, including length, width, height, and available work area. Record each device's dimensions, basic functions, location in the production process, and material flow direction and volume with other equipment. Model the equipment layout problem as finding the optimal placement of equipment within a given workshop space to minimize material transportation costs and maximize equipment collaboration efficiency. Use the coordinate position of each device as a decision variable, and all decision variables form a solution vector, representing a specific equipment layout solution. At the same time, consider constraints such as safe distances between devices and operating space.

[0062] Application of the Cuckoo Search Algorithm: The Cuckoo Search Algorithm simulates the parasitic behavior of cuckoos. In equipment layout optimization, each cuckoo represents an equipment layout solution, and its position represents the solution vector for the equipment layout. The algorithm uses a Levy flight mechanism to conduct a global search to discover new potential layout solutions. Levy flights are a search strategy with random steps that enable cuckoos to make long-distance jumps within the search space, helping them escape local optima. In each iteration, some cuckoos (with a certain probability) perform Levy flights to generate new positions (i.e., new equipment layout solutions). These newly generated layout solutions are then evaluated, and their fitness is calculated based on material transportation costs and equipment coordination efficiency. If the fitness of a new solution is better than the current optimal solution, it is replaced. Furthermore, to maintain population diversity, some poor nests (i.e., poor equipment layout solutions) are randomly discarded with a certain probability and new solutions are generated. Through continuous iteration, a more optimal equipment layout solution is gradually found.

[0063] Optimization plan development and implementation: After a certain number of iterations (e.g., 1,200), the optimal equipment layout plan is determined. For example, equipment with high material flow rates is arranged as close together as possible to reduce material transportation distances; related production equipment is rationally arranged according to the production process sequence to improve equipment collaboration efficiency. During implementation, the company's workshop planning system is used to rearrange equipment according to the optimized layout plan. During this rearrangement process, factors such as equipment installation and commissioning, as well as the planning of workshop logistics channels, are fully considered. Simultaneously, material transportation routes and equipment collaboration are monitored in real time to ensure the effective implementation of the optimization plan.

[0064] Implementation results: Material transportation energy consumption was reduced by 18%, equipment collaborative efficiency increased by 25%, and by optimizing equipment layout, the production logistics of the workshop was effectively improved, and the overall production efficiency and energy utilization rate were improved.

[0065] Example 2:

[0066] Joint Optimization of Production Planning and Inventory Based on the Bat Algorithm: Medium-sized waste plastic recycling companies face the interconnected and complex challenges of production planning and inventory management. Irrational production planning can lead to inventory overstocks or shortages, increasing inventory costs and impacting customer satisfaction. Furthermore, inventory levels can limit the flexibility of production plans. For example, excessive raw material inventory consumes significant capital and storage space, while insufficient inventory can lead to production disruptions. Therefore, an optimization method that comprehensively considers production planning and inventory management is needed to maximize business profitability.

[0067] Data Collection and Problem Modeling: Collect the company's order data, including order quantity, delivery time, product specifications, etc.; production capacity data, such as the production speed of each production equipment and available production time; inventory data, such as the raw material inventory quantity, finished product inventory quantity, safety stock level, etc.; and cost data, such as production cost, inventory holding cost, and stock-out cost. Model the joint optimization problem of production planning and inventory as determining the optimal production plan (including production time and production quantity) and inventory strategy (such as raw material procurement time, procurement quantity, and finished product inventory control strategy) to minimize total cost (including production cost, inventory holding cost, and stock-out cost) while meeting order demand. The relevant parameters of the production plan and inventory strategy are used as decision variables, and all decision variables form a solution vector, representing a joint optimization solution.

[0068] Application of the Bat Algorithm: The Bat Algorithm simulates the hunting behavior of bats using echolocation. In the joint optimization of production planning and inventory, each bat represents a joint optimization solution, and its position represents the solution vector. Bats explore the search space by emitting ultrasonic waves (corresponding to the adjustment of the solution vector) and use echolocation (corresponding to fitness evaluation) to determine the location of prey (corresponding to the optimal solution). In each iteration, the bats adjust their position based on their own speed and the position update formula to generate a new joint optimization solution. The speed update formula considers the bat's current speed, the distance between the current position and the global optimal position, and random perturbations to balance global and local search capabilities. The position update formula determines the new position based on the updated speed. The fitness value, or total cost, of the newly generated solution is calculated. If the fitness of the new solution is better than the current optimal solution, it is replaced. Through continuous iteration, the bat colony gradually finds the optimal joint optimization solution for production planning and inventory.

[0069] Optimization plan development and implementation: After multiple iterations (e.g., 900), the optimal joint optimization plan is determined. For example, production schedules and quantities are rationally scheduled based on order demand forecasts and production capacity. Raw material procurement plans are also optimized to ensure inventory levels meet production needs without creating excessive backlogs. During implementation, production planning and inventory operations are executed according to the optimized plan using the company's production and inventory management systems. Order fulfillment status, inventory level changes, and cost data are monitored in real time, allowing the plan to be dynamically adjusted based on actual conditions.

[0070] Implementation results: Total costs were reduced by 16%, inventory backlogs were reduced by 30%, and the out-of-stock rate was reduced to below 5%. By jointly optimizing production planning and inventory management, the company's economic benefits and operational stability were effectively improved.

[0071] Example 3

[0072] Multi-Objective Production Scheduling Optimization Based on an Artificial Bee Colony Algorithm: Large-scale waste plastic recycling companies must simultaneously consider multiple objectives during production scheduling, such as on-time order delivery, minimizing production costs, minimizing energy consumption, and maximizing equipment utilization. These objectives often conflict with each other. For example, improving equipment utilization may increase production costs or delay order delivery. Traditional production scheduling methods struggle to effectively balance these multiple objectives, leading to indecision-making difficulties for companies and impacting overall production efficiency.

[0073] Data Collection and Problem Modeling: Collect order-related data, such as order quantity, delivery time, and product process requirements; production resource data, including equipment quantity, production capacity, maintenance schedule, and human resource skill levels and working hours; and energy data, such as energy consumption parameters and energy prices for different equipment. The multi-objective production scheduling problem is modeled as finding an optimal production task allocation and scheduling solution to simultaneously optimize multiple objectives, including on-time order delivery, production costs, energy consumption, and equipment utilization. The assignment of each production task (which equipment it is assigned to, who will operate it, and when) is considered a decision variable. All these decision variables form a solution vector, representing a production scheduling solution.

[0074] Application of the Artificial Bee Colony Algorithm: The artificial bee colony algorithm simulates the honey-gathering behavior of bees. In multi-objective production scheduling optimization, food sources represent production scheduling solutions. Employed bees, observer bees, and scout bees each implement a different search strategy. Employed bees, based on their experience, conduct local searches near the current food source to find a better production scheduling solution. Observer bees, based on information relayed by employed bees, select a food source (i.e., a production scheduling solution) with a certain probability for further search. Scout bees randomly search for new food sources within the search space to maintain population diversity. For each food source (production scheduling solution), a fitness value is calculated based on multiple objectives. Multi-objective optimization methods (such as non-dominated sorting and crowding calculation in the non-dominated sorting genetic algorithm NSGA-II) are used to determine its performance within the multi-objective space. Through continuous iteration, the bee colony gradually finds a set of Pareto optimal solutions (i.e., optimal solutions that are not simultaneously dominated by other solutions with respect to multiple objectives). From this set of Pareto optimal solutions, enterprises can then select the most appropriate production scheduling solution based on their strategic priorities.

[0075] Optimization plan development and implementation: After multiple iterations (e.g., 1,500), a set of Pareto-optimal solutions is obtained. For example, based on current market demand and its own cost-control strategies, the company selects a Pareto-optimal solution that prioritizes on-time order delivery and minimizes energy consumption. During implementation, the company's production scheduling system is used to schedule production tasks according to the selected solution and monitor the achievement of each objective in real time. If unexpected situations such as order changes or equipment failures occur during production, the artificial bee colony algorithm is rerun to find a new optimal production scheduling solution based on the current state.

[0076] Implementation results: The on-time delivery rate of orders increased to over 95%, energy consumption decreased by 14%, and equipment utilization increased by 20%. A good balance was achieved among multiple goals, improving the company's overall production efficiency.

[0077] The newly added examples above provide a richer range of ideas and methods for optimizing waste plastic recycling production processes from various perspectives, including equipment layout, joint optimization of production planning and inventory, and multi-objective production scheduling optimization. Based on your actual needs, you can provide further optimization suggestions or supplementary information for these examples to better meet your specific production scenarios and business development needs.

Claims

1. A method for scheduling solid waste plastic production tasks and balancing resource allocation, characterized in that: include: Data collection: Collect data on production operations, energy consumption, quality impact, equipment maintenance, production environment and market demand; Problem modeling: Under the premise of satisfying quality, link relationship, equipment maintenance and market demand, with operation selection, time adjustment and market response strategy as decision variables, a model for finding the optimal production sequence is constructed, with the goal of minimum energy consumption, maximum efficiency and market response; let the production sequence solution vector be S = [s1, s2, ..., s n ], the quality of the solution is evaluated by the objective function Z = w1E(S0+w2P(S)+w3M(S), where E(S) is the energy consumption function, P(S) is the production efficiency function, M(S) is the market response function, and w1, w2, and w3 are weight coefficients, which are determined according to the enterprise's production goals and market conditions; Algorithm implementation: Initialize the solution and taboo table, select the best solution from the neighborhood during iteration, and generate neighborhood solutions by changing the operation type, sequence, number of times, operation conditions or market strategy; lift the taboo according to conditions, and dynamically maintain the taboo table according to solution quality, market stability, etc.; Implementation of the plan: Dynamically determine the number of iterations based on factors such as production scale to obtain the optimal sequence, arrange production in combination with production management and market forecasting systems, monitor in real time and adjust according to the market.

2. The method according to claim 1, characterized in that: The taboo list has a dynamic deadline based on solution quality, market and environmental changes, and will be removed based on re-evaluation criteria when it expires.

3. The method according to claim 1, characterized in that: The generation of neighborhood solutions takes into account the impact on the previous and subsequent links, equipment, environment and market, including changes in operation types, sequence exchanges, number of adjustments and market strategy adjustments.

4. A method for optimizing the allocation of waste plastic recycling resources based on the whale optimization algorithm, characterized in that: include: Data collection and problem definition: Collect data related to resource demand, cost, loss, suppliers and supply chain, and define the resource allocation problem that pursues the optimal efficiency, energy consumption, cost and supply chain collaboration under resource and supply chain constraints. Algorithm application: The whale position represents the resource allocation plan, which is adjusted according to the encirclement, bubble net attack and prey search behavior. In the spiral update position of the bubble net attack phase, the whale position update formula is X i (t+1)=X * (t)+D·e bl ·cos(2πl), where X i (t+1) is the position of the i-th whale at time t+1, X * (t) is the current optimal solution position, D = |C·X * (t)-X i (t)|, C is a random vector, b is a constant that controls the shape of the spiral, and l is a random number between (-1,1). The adjustment parameters at each stage change dynamically according to factors such as resources, production, cost, and supply chain, and the search range and step size are adaptively adjusted according to population, market, and supply chain risks. Solution implementation: Determine the optimal solution with the number of iterations based on factors such as resources and supply chain, use intelligent resource management and supply chain collaboration systems to allocate resources, monitor in real time, and make adjustments based on supply chain changes.

5. The method according to claim 4, characterized in that: In the encirclement phase, whales move toward the optimal solution according to the proportions related to resources, suppliers, and supply chains. The proportions are dynamically adjusted based on iterations, resources, costs, link priorities, and supply chain risks.

6. The method according to claim 4, characterized in that: During the bubble network attack phase, the contraction amplitude and spiral parameters change dynamically depending on resources, production, cost, supply chain and iteration coefficient.

7. A method for optimizing the formula of recycled waste plastic products based on differential evolution algorithm, characterized in that: include: Data collection and model building: Collect data on raw materials, product performance, cost, energy consumption, market and competition, and build multiple models to describe the relationship between formula and various factors. Algorithm application: Take the raw material ratio as the decision variable and optimize the formula through mutation, crossover and selection operations. In the mutation operation, the formula for generating mutant individuals is V i，G =X r1，G +F×(X r2，G -X r3，G ), where V i，G is the i-th mutant individual of the G-th generation, X r1，G , X r2，G , X r3，G are three different individuals randomly selected from the Gth generation population, and F is the mutation factor. Each operation is adaptively adjusted based on factors such as raw materials, performance, market, and competition, and the fitness function integrates factors such as product quality, cost, energy consumption, market response, and competitive advantage. Implementation of the plan: Determine the optimal formula through the number of iterations based on factors such as products, markets, and competition, utilize a variety of monitoring and intelligence systems to produce, and fine-tune based on actual conditions, markets, competition, and environmental policies.

8. The method according to claim 7, characterized in that: The mutation factors are adjusted dynamically based on population, raw materials, performance, market and competition, enhancing global search in the early stages of evolution and local search in the later stages.

9. The method according to claim 7, characterized in that: Crossover factors are dynamically adjusted based on individual fitness, performance, market, competition and environmental requirements to determine whether genes are retained or updated.

10. The method according to claim 7, characterized in that: The fitness function integrates product quality, market demand, competitive advantage, production cost, energy consumption, market penalty and environmental cost. The weight and calculation method of each score are determined according to the corresponding factors and models.

Images