Parallelized Swarm Intelligence Approach for Solving TSP and JSSP Problems

Abstract

One of the possible approaches to solving difficult optimization problems is applying population-based metaheuristics. Among such metaheuristics, there is a special class where searching for the best solution is based on the collective behavior of decentralized, self-organized agents. This study proposes an approach in which a swarm of agents tries to improve solutions from the population of solutions. The process is carried out in parallel threads. The proposed algorithm—based on the mushroom-picking metaphor—was implemented using Scala in an Apache Spark environment. An extended computational experiment shows how introducing a combination of simple optimization agents and increasing the number of threads may improve the results obtained by the model in the case of TSP and JSSP problems.

Mushroom Picking Framework with Cache Memories for Solving Job Shop Scheduling Problem

“…Mushroom Picking Algorithm was introduced in [30] for solving difficult optimization problems. It is characterized by the following features:…”

“…Another metaheuristic designed for parallel environments named Mushroom Picking Algorithm (MPA) was proposed by the authors in [30]. The approach proved successful in solving the JSSP instances.…”

An optimization algorithm for TSP-like problem in warehouse picking

“…It is easy to see from figures 1 and 2 that if the starting point and the ending point are the same, it is a classic TSP (travel agent) problem, but now the starting point and the ending point may or may not be the same. This paper calls it a TSP-like problem, and it is obvious that the TSP-like problem is a special case of the TSP-like problem [1].…”