The research of dynamic plant layout problem (DPLP) is to find the optimal layout scheme within multiple periods of production, with the increasing scale of calculation, the efficiency of the traditional centralized algorithm may be decreased. In this paper, we proposed a decentralized genetic algorithm (DGA) which combines genetic algorithm (GA) with decentralized optimization technology. Simulation experiments are carried out to compare the decentralized genetic algorithm with the centralized one, and the result shows that the decentralized method is better than the latter at convergence rate and optimization result. On the other way, the decentralized algorithm is a parallel computing technology, so it can improve computational efficiency especially in high-complexity systems.

No abstract

No abstract