# A hybrid CPU-GPU parallelization scheme of variable neighborhood search for inventory optimization problems

## TL;DR

This paper explores hybrid CPU-GPU parallelization strategies for the Variable Neighborhood Search metaheuristic, demonstrating promising computational improvements on complex, NP-hard inventory optimization problems involving reverse logistics.

## Contribution

It introduces a hybrid parallelization scheme for VNS on CPU-GPU systems and evaluates its effectiveness on challenging inventory optimization benchmarks.

## Key findings

- Hybrid parallel VNS improves computational efficiency.
- Parallelization approaches show promising results.
- Effective for NP-hard inventory problems.

## Abstract

In this paper, we study various parallelization schemes for the Variable Neighborhood Search (VNS) metaheuristic on a CPU-GPU system via OpenMP and OpenACC. A hybrid parallel VNS method is applied to recent benchmark problem instances for the multi-product dynamic lot sizing problem with product returns and recovery, which appears in reverse logistics and is known to be NP-hard. We report our findings regarding these parallelization approaches and present promising computational results.