Research On Scheduling Optimization Of PVC Production Based On Estimation Of Distribution Algorithm

Polyvinyl chloride(PVC)is a widely used chemical product.Most of our country uses the calcium carbide method to produce polyvinyl chloride.It has the characteristics of multiple processes,multiple equipment,and high energy consumption.Therefore,enterprises need to optimize their production scheduling to achieve the purpose of energy saving,consumption reduction and production cost reduction.In order to shorten the order completion time and improve production efficiency,a production scheduling model with the goal of minimizing the maximum completion time is established based on the production characteristics of the aggregate section.Because the model is large and difficult to solve,an estimation of distribution algorithm is proposed to solve it.In order to further improve the efficiency and quality of the estimation of distribution algorithm,the improvement strategy of the estimation of distribution algorithm is studied.In addition,considering that there are often multiple production objectives in the actual production process,the multi-objective and multi-period scheduling optimization problem of the whole process of PVC production is studied.The main researches in this thesis are as follows:Aiming at the scheduling problem of aggregation section,a production scheduling mathematical model is established with the goal of minimizing the maximum completion time.The coding method based on machining task is designed,and the distribution estimation algorithm is used to solve the model,which improves the solving efficiency of the algorithm.The estimation of distribution algorithm uses random initialization of the population to ensure the dispersion and diversity of the population;by sorting the fitness values of the population individuals,selecting the dominant group to construct a probability model,it can accurately describe the distribution information of the dominant solution;generating a new population through the probability model to achieve population evolution.Finally,through simulation experiments of different examples and comparison with genetic algorithm and particle swarm algorithm,it is verified that the estimation of distribution algorithm has its advantages in terms of the quality of the optimized solution and the speed of convergence.In order to obtain the optimal scheduling scheme and improve the efficiency of the solution,the improvement strategy of the estimation of distribution algorithm is studied.Two encoding methods based on aggregation order-feeding times and aggregation matrix are designed,which can reduce the complexity of encoding and decoding and increase the diversity of task arrangements.The convergence speed and precision of the algorithm can be improved by integrating proportion increasing strategy and binary search strategy into the algorithm.Finally,the experiment is compared to verify that the solving efficiency of the improved estimation of distribution algorithm and the quality of the optimal solution are significantly improved,and the optimal scheduling scheme can be obtained.Aiming at the problem of multiple production indicators in the actual production process,the whole process of PVC production is the research object,and the optimization goal is to minimize the maximum completion time and minimize the production cost,and establish a multi-objective and multi-period scheduling optimization model for the whole process of PVC production by calcium carbide method.The weighted sum method and improved distribution estimation algorithm are used to solve the model and obtain the optimal scheduling scheme.By comparing the solution results with the single-objective model,it can be seen that the multi-objective and multi-period model can achieve on-time delivery and reduce production costs.It verifies the effectiveness of the multi-objective and multi-period scheduling optimization model in actual production,and it can guide the production scheduling of the PVC factory.