Simultaneous allocation of buffer capacities and service times in unreliable production lines

Simultaneous allocation of service times and buffer capacities in manufacturing systems a random environment is NP-hard combinatorial optimisation problem. This paper presents sophisticated simulation-based approach for the design unreliable production lines to maximise rate. The proposed method allows global search using Genetic Algorithm (GA), which coupled with Finite Perturbation Analysis (FPA) as local technique. Traditional techniques based on perturbation analysis optimise decision variables same nature (e.g. time only, capacity only), whereas technique simultaneously provides an capacities. One main focuses this investigation persistence or absence rate patterns are among most essential insights that come from designing lines. results show superiority combined GA-FPA regarding GA FPA terms solution quality convergence behaviour. Moreover, considering instances ranging 3 100 machines, our numerical experiments line literature small (as similar identified work), but important differences highlighted medium/large instances.