A decomposition-based algorithm for the double row layout problem

Abstract

The double row layout problem (DRLP) is a common pattern of facility layout problem, which has practical applications in flexible manufacturing systems. The double row layout problem is vital to save transportation cost and enhance productivity. Nevertheless, it is very hard to handle the DRLP because of its characteristic of combination of combinatorial and continuous aspects. In this paper, a decomposition-based algorithm is proposed to solve the DRLP. We decompose the DRLP into two subproblems. In the first subproblem, the adjustable clearances between adjacent facilities are temporarily ignored. A first improvement based local search is applied to optimize the sequences of facilities on double rows. During this process, the facilities of double rows are placed starting at different abscissas rather than starting at the same abscissa for each arrangement. A property of the objective function of the DRLP is used to obtain the optimal difference between two starting abscissas. In the second subproblem, a particle swarm optimization is applied to optimize the adjustable clearances between adjacent facilities under the condition that the sequences of facilities are fixed. Our proposed algorithm is evaluated on 59 test instances and compared with the state-of-the-art methods. The experimental results demonstrate the high competitiveness of our proposed algorithm.