Defining Parallel Local Search Procedures with Neighborhood Combinators

This paper presents a declarative approach for building parallel local search algorithms. The goal is to easily achieve speed improvements thanks to the growth both in multi-core hardware and the massive availability of distributed computing power, notably in the cloud. Local search algorithms rely on the exploration of neighborhoods on a given solution space according to a problem model. Our approach relies on the exploration of multiple neighborhoods in parallel, performed by different workers that can be located on different CPU cores (locally or remotely). This approach is based on neighborhood combinators, which are composite neighborhoods built out of basic ones. Combinators are a domain-specific language to build local search procedures out of building blocks such as metaheuristics, neighborhood selection, stop criterion and all other relevant aspects commonly found in local search procedures. This paper proposes a set of combinators that introduce parallel optimization. An implementation is included in the OscaR.cbls framework, using the Akka Actor model of computation. © 2022, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.