Study On Unrelated Parallel Machines Scheduling Problem With Shelf-life Raw Materials Constraints

With safety accidents of foods and pharmaceuticals frequently occurring due to expired raw materials used in production process,the safe use of raw materials has aroused widespread concern in the recent years.Because raw materials are the objects of production and processing,it is necessary that raw materials are safely used during production and processing in order to ensure a safe and smooth production process.However,owing to the natures of raw materials and the external storage environments and other factors,raw materials easily expire and deteriorate once it is opened in the process of production,which brings a big challenge for production and scheduling.In practice,most of production and processing like foods and pharmaceuticals,are implemented under unrelated parallel-machine scheduling environment.Therefore,based on these thinking,the production scheduling problem with shelf-life raw materials under the environment of unrelated parallel machines will be discussed in this thesis.First,the thesis focuses on usage states of raw materials and further analyze effects of production decisions under different states on subsequent productions,in order avoid the interruption of production process by raw materials expired or stock out.Secondly,the thesis further studies a single-objective unrelated parallel machines scheduling problem with shelf-life raw materials constraints.A non-linear mixed integer programming model is formulated to minimize total production costs which is the total combination of raw material costs and machines operating costs.In view of the NP-hardness of this problem,Variable Neighborhood Discrete Particle Swarm Optimization algorithm is proposed,inheriting the thought of particle swarm optimization algorithm to guarantee rapid evolution speed,and redefining the operations of particle swarm optimization algorithm based on the nature of the problem,to overcome the shortcoming of premature convergence.Then,the algorithm hybridizes variable neighborhood search algorithm,which can improve the diversity of solutions set.Computational experiments have been conducted to demonstrate the effectiveness and superiority of Variable Neighborhood Discrete Particle Swarm Optimization algorithmFinally,bi-objective unrelated parallel machines scheduling problem with shelf-life raw materials constraints is proposed to minimize the total completion time and raw material costs and then a bi-objective non-linear mixed integer programming model is formulated.Because the problem not only is NP-hard problem,but also is a bi-objective optimization problem,Evolutionary Discrete Particle Swarm Optimization algorithm is proposed.This algorithm based on particle swarm optimization algorithm defines hybrid-greedy initialization method,shortest-processing-time local search rule and other operators,to ensure the quality and diversity of solutions The effectiveness and superiority of overall algorithm and operators' behavior are verified by computational experiments.The research enriches theories of production scheduling by proposing new production scheduling models and new intelligent optimization solving methods.Moreover,the results provide theoretically decision supporting for production scheduling problems with shelf-life constraints.In practice,this research is more feasible and tractable,which is helpful to managers in making effective operation decision.