Analysis of simulated annealing cooling schemas for design of optimal flexible layout under uncertain dynamic product demand

Manufacturing facilities are subjected to many uncertainties such as variability in demand, queuing delays, variable task times, rejects and machine breakdown. These volatilities have a large impact on leap time, inventory cost and delivery performance of a manufacturing unit. To operate efficiently the manufacturing facilities should adapt to these variations. The paper explores the way uncertainties are addressed in designing of flexible optimal layout. Such facility layout problem is known as stochastic dynamic facility layout problem (SDFLP). SDFLP is an NP-hard combinatorial optimization problem, which means the time taken to solve increases exponentially with problem size. To solve SDFLP, the paper presents an adaptation of simulated annealing (SA) meta-heuristic. Various SA cooling schemas are discussed, computed and evaluated for generating the optimal flexible layout. An optimal layout is one that minimises the distance travelled by materials taking into account uncertain product demand (material handling cost). A computer-based tool was developed and analysis was conducted on small to large size problem set. The results showed that SA with exponential cooling schedule provides better solution in terms of layout efficiency and gave better solution as compared to literature.