Optimizing a realistic large-scale frequency assignment problem using a new parallel evolutionary approach

Abstract

International audienceThis paper analyzes the use of a novel parallel evolutionary strategy to solve complex optimization problems. The work developed here has been focused on a relevant real-world problem from the telecommunication domain to verify the effectiveness of our approach. The problem, known as frequency assignment problem (FAP), basically consists of assigning a very small number of frequencies to a very large set of transceivers used in a cellular phone network. Real FAP instances are very difficult to solve due to the NP-hard nature of the problem, therefore using an efficient parallel approach which takes the most of different evolutionary strategies can be considered as a good way to obtain high-quality solutions in short periods of time. Specifically, a parallel hyper-heuristic based on several meta-heuristics has been developed. After a complete experimental evaluation, results prove that our approach obtains very high-quality solutions for the FAP (it beats any other result published)

Parallel computational optimization in operations research: A new integrative framework, literature review and research directions

“…We found four studies which parallelize heuristics that differ from all types described above: an agent-based heuristic [Benedicic et al, 2014], an auction-based heuristic [Sathe et al, 2012], a Monte Carlo simulation inside a heuristic-randomization process [Juan et al, 2013], and a random search algorithm [Sancı andİşler, 2011]. We found two studies which parallelize matheuristics [Stanojevic et al, 2015, Groer et al, 2011 and three studies which suggest multi-search algorithms [Chaves-Gonzalez et al, 2011, Vidal et al, 2017, Lahrichi et al, 2015. Due to the diverse nature of the aforementioned studies, we do not look for patterns in algorithmic parallelization, computational parallelization and scalability results.…”

A novel parallel multi-swarm algorithm based on comprehensive learning particle swarm optimization

“…In Li and Wada (2011), Li and Wada proposed a parallel PSO algorithm which reduced effectively the communication latency and improved the effectiveness of PSO on distributed environment. In Chaves-González et al (2011), a parallel hyperheuristic based on seven heuristics was proposed in order to solve complex optimization problems. To prove the performance of this algorithm, this approach was tested on a real-world problem in telecommunication.…”