Research On The Multiobjective Cooporation Optimazation Of Mixed-Model Flow Manufacturing System

This dissertation mainly investigates the problem of multi-objective cooperative optimization on the design of the mixed-model flow manufacturing. Based on the cooperation design ideology, the cooperative and mutual effect of balancing and scheduling of mixed-model flow line is taken into consideration in the design, thus, the design is balancing-to-scheduling and scheduling-to-balancing. The co-evolution algorithm is applied to the optimization problem, and a new optimization system and algorithm frame of designing of mixed model flow manufacturing is established. In the algorithm frame, the co-evolution algorithm combing with the niche technique, the selection mode of face-to-rank and Pareto classifier, makes the solution of system be global optimal, not local optimal; and the single objective optimal solution is extended to multiple objective Pareto optimal solution. Based on result of the multiple objective optimization, this dissertation investigates the multiple objective decision of mixed-model flow manufacturing system. In allusion to the Pareto frontier, in terms to the assistant information of decision schemes, the preference of objectives, the preference of decision schemes, individual decision-making and group decision-making are gained through the measure function, 0-1 programming and relative entropy combining with subjective and objective factors.Based on the optimization system and algorithm frame, the co-evolution optimization model of minimization of makespan, minimization of workload variation, leveling of the material flow and U-line is established, and some instances is given also. In the study of the multi-objective cooperative optimization on makespan, the relationship between various balancing indices with the system objective is analyzed; the result of co-evolution optimization and result of serial optimization is compared and analyzed; the algorithm is given. Based on the decision schemes of co-evolution, the multi-objective decision is made. In the model of leveling the material flow, the model of minimizing usage rate is extended to all levels of materials in the flow line, and a new model is built up, based on the makespan; the objective of leveling the material flow is combined with the line balancing firstly. In addition, according to difference between the top-level materials and the low-level materials, the model of leveling the material flow is set up separately. Furthermore, the model of leveling the material flow with constrain of resource supplying limit. In the model of optimization on the workload, the mutual effect of line balancing and schedulingis analyzed by an instance; the factors of line balancing, scheduling, and system parameters, such as workstation length, the launch rate are taken into considered. Four models under the combination two type of work station with two type of hypotheses, and an instance is given. In model of the optimization of U-line, according to the characteristic of U-line, a model of optimization on the workload is set up, the optimization methods is given. Moreover, a method of workstation breaking is put forward, thus, the problem of optimization of U-line can be transferred into straight-line problem and the two problems are unified through the method.