[2J4-GS-1] Fundamental AI, theory: algorithm

[2J4-GS-1-03] Speeding Up Solving Large-Scale Vehicle Routing Problems Using Hybrid Quantum-classical Computation

〇Eiji Kawase¹, Hideaki Tamai¹ (1. Oki Electric Industry Co., Ltd.)

Keywords:Vehicle Routing Problems, Hybrid Quantum-classical Computation, Logistics

In this study, we perform a quantum-classical hybrid computation using a classical computer and a quantum annealing machine in order to speed up the solution of large-scale vehicle routing problems. We describe the results of using a quantum annealing machine to solve the maximum cut problem, dividing the bases into several areas, and solving the optimal routing plan for each area using a mathematical optimization solver on a classical computer. As a result, the quantum-classical hybrid calculation was 26 times faster than the conventional heuristic calculation alone, and the accuracy of the initial solution was improved by 63%. Compared to area dividing of using classical k-means, the initial solution was obtained about 1.6 times faster and with the same accuracy.