MODELING A COMPLEX REAL-WORLD MANUFACTURING SCHEDULING PROBLEM

We present a real-world manufacturing setting that consists of many features, each of which adds complexity to the underlying scheduling problem. The problem is composed of identical jobs. A partially ordered set of operations must be applied to each job. Each operation requires a resource, in the form of a laboratory in which the dedicated machine resides. Each operation is processed by an employee, chosen from a subset of employees eligible to perform this operation. Additional features include set-up times between laboratories, capacity-restricted queues before/after an operation, varying processing times of the different employees, and a requirement to re-process an operation due to stochastic manufacturing failures. The problem is related to several existing NP-hard scheduling problems – reentrant flow shop scheduling, flexible shop scheduling, and cyclic shop scheduling.