Development and validation of genetic algorithm-based facility layout – a case study in the pharmaceutial industry

In the layout design of pharmaceutical factories, several important objectives often need to be considered, such as operation cost, maintenance cost, material handling cost, throughput rate, etc. Most of the existing algorithms for facility layout design were developed based on pre-determined single objectives; namely, the distance-based objectives or the adjacency-based objectives. In this paper, we propose a genetic algorithm approach with an embedded simulation model which allows the user to select the objectives that are important in each particular layout design in the pharmaceutical industry. To verify the feasibility of the proposed method, the layout designs of two pharmaceutical plants for solid dosage forms were used as a test bed. One represents a typical small-size pharmaceutical plant and the other represents a medium-size plant. The objectives are to maximize throughput rate and to minimize travelling time per trip. A simulation model was developed, and the efficiency of the layouts generated by human designers CORELAP, CRAFT, and BLOCPLAN, is compared. The comparison indicates that the proposed genetic algorithm method significantly outperforms human designers and other computer algorithms in minimizing travelling time per trip under the same throughput rate.