Modified genetic algorithms for solving facility layout problems

Abstract

A facility layout design is one of the most commonly faced problems in the manufacturing sectors. The problem is mixed-integer in nature and usually an NP-hard problem. Due to mixed-integer nature of the problem, it is difficult to solve the problem using classical optimization techniques. The classical optimization techniques are better for local search of the optimal solution. However, these algorithms are not efficient when there are multiple optimal solutions and alternate optimal solutions. To overcome these limitations, this paper proposed a new interactive evolutionary algorithm based local search algorithm for solving static facility layout problems with unequal compartments. This is an iterative based two steps algorithm. The evolutionary algorithm creates the new solutions for the local search algorithm to obtain a local optimal solution. The designer can interact between these processes to derive the best possible solution of the problem. The objective function of the problem is non-linear one in which the sum of the material handling cost has been minimized. Apart from the conventional evolutionary operators, i.e. selection, crossover, mutation and elitism, this paper has also proposed exchange and rotation operators. The rotation operator is used to avoid mixed-integer formulation of the problem for the local search problem. The use of rotation operator has also reduced the number of variables of the problem significantly. The performance of the model is tested over previously solved problems selected from the literature. The evaluation of the results shows that the performance of the proposed model is better than many existing algorithms and has the potential for real-world applications.