A Multi-Threaded Cuckoo Search Algorithm for the Capacitated Vehicle Routing Problem

Abstract

Cuckoo search is a bio-inspired algorithm based on the reproduction behavior of some cuckoo species. This metaheuristics seems promising to solve the capacitated vehicle routing problem. This paper analyzes the standard capacitated vehicle routing problem because the cuckoo search enables faster results with fewer parameters than other optimization algorithms. A new approach using a multi-threaded variant of the cuckoo search running on multiple CPU cores is being investigated, which allows the parallelization of optimization cycles. The approach uses a standard Java framework and takes into account multiple eggs per nest. A quantitative analysis investigates the new multi-threading variant compared to the standard one.