Machine layout in 2-dimensional plane based on genetic algorithms

Optimal solution of machine layout problem is very difficult due to non-polynomial nature of the problem. Further, the problem becomes more critical when pickup/drop-off locations are considered in the design of layout. Deterministic techniques are not computationally feasible. In this paper Genetic algorithm (GA), a relatively new heuristic that works according to the biological production process, is used for obtaining efficient layouts with near optimal solutions. Layout of machines is modeled in gene structure. The performance of different genetic algorithm parameters involved in the design of machine layout with accurate positioning of pickup/drop-off locations along the periphery of the rectangular machine blocks is discussed in detail. The algorithm is coded in turbo C language and implemented in a personal computer Pentium III, 550 MHz. The preliminary experimentations were carried out on a 6-machine problem in order to fix the genetic parameters like crossover operator, crossover rate, and population size. The results of preliminary experimentation were utilized to generate the facility layout design for a 12-machine problem. The results obtained through the present investigation depict the fact that a cost and performance effective machine layout can be designed with proper setting of GA operators. Moreover, the proposed technique is simple and more realistic to reach a near optimal solution.