A novel nonuniform discrete time formulation for short-term scheduling of batch and continuous processes

Abstract

We propose a novel nonuniform time discretization method (NUDM) formulation for a shortterm scheduling problem of batch and continuous operations. It aims to reduce the size of resultant MILP models by avoiding the use of a high dimensioned task-time matching binary variable defined in previous NUDM approaches. We devised several efficient constraints, with subdivided binary variables assigned to the start, process, and end event of each task. Several types of examples show that the proposed formulation is both general and efficient. For instance, we discuss fixed and variable processing time problems, and a sequence-dependent setup-time problem. A simple optimization algorithm tailored for a periodic scheduling problem is also proposed.