Multiobjective optimization of buffer capacity allocation in multiproduct unreliable production lines using improved adaptive NSGA-II algorithm

Abstract

Buffer capacity allocation occupies an essential position in the designing of production systems. By means of polymorphism analysis on production capacity and capability, the buffer allocation for multi-stage production lines with unreliable machines, which simultaneously involves the maximization of theoretical production rate in system and minimization of system state entropy for a certain amount of buffers, was investigated in this paper. For repairable modular machines, the Markov models are established using stochastic process analysis and the corresponding theoretical steady-state probability in various states are obtained. Furthermore, the original system in combination with multi-state reliability measures of buffer stations is equivalent to a system with independent machines which can be expressed by vector u-functions. Based on the probability distributions of the states of subsystems, the composition operators for series connections and parallel connections are defined. Consequently, the entire system is simplified to one component represented by the polynomial UGF and the systematic multi-state reliability and structural complexity were assessed. Based on the theoretical production rate in system and system state entropy, a mathematical model for buffer capacity optimization was established and optimized using a specific genetic algorithm. The feasibility and effectiveness of the proposed method was verified by applying which to a production line of engine heads.