Heuristics for flow shop rescheduling with mixed blocking constraints

In the flow shop rescheduling literature, many papers consider unlimited buffer capacities between successive machines. In real fact, these capacities may be limited, or no store may exist. Thus, a blocking situation is inducted. Diverse types of blocking constraints are studied in the flow shop scheduling problems. However, in dynamic environments, only few papers deal with these kinds of constraints. The aim of this paper is to investigate a problem of rescheduling the jobs in a flowshop environment and mixed blocking as a constraint, considering simultaneously schedule efficiency and stability as a performance measure, and job arrival as a disruption. An iterative methodology based on the predictive–reactive strategy is implemented for dealing with this rescheduling problem. The problem has first been modeled as a Mixed Integer Linear Programing (MILP) model. Experimental results show that the MILP resolution is only possible for small-sized instances. Hence, inspired by NEH algorithm, we proposed four heuristics for solving large-sized instances of this problem. Eventually, we discussed the performance of the proposed heuristics for different blocking situations, both in terms of solution efficiency and resolution time.