A computational ecosystem for optimization: review and perspectives for future research

Abstract

Nature exhibits extremely diverse, dynamic, robust, complex and fascinating phenomena and, since long ago, it has been a great source of inspiration for solving hard and complex problems in computer science. Hence, the search for plausible biologically inspired ideas, models and computational paradigms always drew the interest of computer scientists. It is worth mentioning that most bio-inspired algorithms only focuses on and took inspiration from specific aspects of the natural phenomena. However, in nature, biological systems are interlinked to each other, e.g., biological ecosystems. The ecosystem as a whole can be composed by species that respond to environmental and ecological stimuli. This work reviews the theoretical foundations and applications of a computational ecosystem for optimization, named ECO. Also, as some concepts and processes inherent to biological ecosystems have already been explored in the ECO approach, some related works are described. Finally, several future research directions are pointed.

Evolutionary computation and swarm intelligence for the inference of gene regulatory networks

“…Anyhow, the overall analysis of results suggests that DE is the most promising algorithm for the problem, since it consistently achieved good results. Therefore, improvements in the DE algorithm can be a starting point for future work, with special attention to hybridisation, since it has been shown very efficient for hard optimisation problems (Parpinelli and Lopes, 2015). Also, further work will include more experiments regarding to the sensitivity of the algorithms to changes in the parameters of the GRNs, such as the number of genes in the network, the average number of connections k and the number of temporal expressions.…”

A discretized differential evolution algorithm for the inference of Gene Regulatory Networks

“…Further research direction include improvements in the evolutionary algorithm and the use hybrid methods, such as those proposed by [27]. Also, variations of the fitness function will be considered, such as by using the Tsallis Entropy, as shown by [28].…”

An efficient multi-core implementation of the Jaya optimisation algorithm

“…In [27], they implemented a multi-population model based parallel ABC algorithm in MPI and investigated an extended performance study about different migration parameters and communication topologies among the subpopulations. Recently, a hierarchical cooperative search scheme has been reported inspired by the natural ecosystem [28]. The ecosystem as a whole can be composed by populations and each population running a metaheuristic algorithm.…”

“…By this way, there are two levels of communications, intra-habitats and inter-habitats communication, that favors co-evolution. These two communications can be implemented through ecological symbiotic relationships [28]. The advantage of this ecosystem is that there is a better balance between exploration and exploitation ability and making the algorithm perform well in complex optimisation problems.…”