A Variable Neighborhood Search (VNS) metaheuristic for Multiprocessor Scheduling Problem with Communication Delays

Abstract

The purpose of this paper is to present a Variable Neighborhood Search (VNS) metaheuristic for solving the Multiprocessor Scheduling Problem with Communication Delays (MSPCD). The MSPCD problem considers scheduling task graph on a multiprocessor system, taking into account communication delays. The task graph contains precedence relations as well as the amount of exchanged data between tasks. The multiprocessor architecture is composed by a set of identical processors connected in an arbitrary way. The communication is defined throw two symmetric matrices containing, respectively, the communication rate and the access cost between each two processors. Unlike the proposed VNS in the literature, we use a different representation of the solutions that leads to the definition of two natural neigh-borhoods structures. Computational experience shows advantages of our approach.