A Parallel Large Neighborhood Search-Based Heuristic for the Disjunctively Constrained Knapsack Problem

This paper proposes a parallel large neighborhood search-based heuristic for solving the Disjunctively Constrained Knapsack Problem (DCKP), which has an important impact on the transportation issues. The proposed approach is designed using Message Passing Interface (MPI). The effectiveness of MPI´s allows us to build a flexible message passing model of parallel programming. Meanwhile, large neighborhood search heuristic is introduced in the model in order to propose an efficient resolution method yielding high quality solutions. The results provided by the proposed method are compared to those reached by the Cplex solver and to those obtained by one of the best methods of the literature. As shown from the experimental results, the proposed model is able to provide high quality solutions with fast runtime on most cases of the benchmark literature.