Bio-inspired optimization algorithms

Bio-inspired optimization algorithms to efficiently solve optimization problems in diverse application areas.

Bio-inspired optimization algorithms are those methods that are generally inspired by physical principles, evolution theory and certain behaviors of living beings to efficiently solve optimization problems in very diverse application areas [1]. The history of these algorithms begins in 1975 when John Holland et al. proposed the first optimization model based on Charles Darwin’s theory of evolution, a model that received the name of Genetic Algorithm (GA) [2]. Thanks to the possibilities offered by genetic algorithms compared to traditional optimization methods and with the increase in the computational capacity of technology, a large number of bio-inspired algorithms were developed from the year 2000 onwards. Among these optimization methods we can find from algorithms that simulate the alignment of particles during the cooling of materials (Simulated Annealing) to algorithms that model the movement of galaxies (Galaxy-based Search Algorithm), to algorithms based on the collective behavior of certain species of animals [1, 3]. Some of the most important algorithms are shown in Fig. 1.
Bio-inspired optimization algorithms examples
Fig. 1: Examples of bio-inspired optimization algorithms.
Although every bio-inspired optimization algorithm has been designed in a generic way to deal with any optimization problem, each bio-inspired algorithm has its own particularities that make it a better choice over other algorithms to solve a specific problem. For example, ant colony algorithms are a good choice for dealing with routing problems, genetic algorithms are a good choice in a general way as long as complex solution reconstruction methods do not have to be developed when obtaining infeasible solutions after applying evolutionary operators, and the bee colony algorithm is usually a good choice for dealing with combinatorial optimization problems. It is therefore up to the expertise of the expert to choose the correct algorithm to apply for each problem, as well as the design of the switching functions and solution optimization that best fit that problem.

Why are bio-inspired optimization algorithms interesting compared to other optimization methods?

Originally, bio-inspired optimization algorithms were proposed as methods to deal with very complex optimization problems whose size made the use of exact solution methods unfeasible due to the amount of time required to obtain the optimal solution. In the literature, these problems are called NP-complete problems. The underlying idea for the use of these algorithms, which is the reason why these algorithms are really interesting, lies in taking advantage of the laws of physics and the mechanisms that nature itself has designed and improved over millions of years to solve the three major obstacles that a living being can face: reproduction, obtaining food and adapting to the environment. With these algorithms we can obtain quasi-optimal solutions to complex problems in a reasonable time. In addition, many of the bio-inspired optimization algorithms are based on evolving populations of individuals, each individual being a concrete solution to the problem. The latter makes the methods of these algorithms very easy to parallelize, allowing us to considerably reduce the time to obtain a near-optimal solution by taking advantage of the benefits of parallel computing on one or even several machines. BIBLIOGRAPHY [1] A. Darwish, “Bio-inspired computing: Algorithms review, deep analysis, and the scope of applications,” Future Computing and Informatics Journal, vol. 3, no. 2, pp. 231-246, 2018. [2] Xin-She Yang, “Genetic Algorithms,” Nature-Inspired Optimization Algorithms, chapter 6, 2nd edition, Academic Press, pp. 91-100, 2021. [3] M. A. Rahman et al., “Nature-Inspired Metaheuristic Techniques for Combinatorial Optimization Problems: Overview and Recent Advances,” Mathematics, vol. 9, no. 20, 2633, 2021.

Acerca del autor

Víctor Berrocal oga
Víctor Berrocal
Senior Optimization Consultant en oga