Nature-Inspired Evolutionary Computation For Optimization

Volume 12 Issue 2 October - December 2017

Research Paper

Nature-Inspired Evolutionary Computation For Optimization

B.V. Babu*

Former Vice Chancellor, Graphic Era University, Dehradun, India & Galgotias University, Uttar Pradesh, India and Former Professor & Dean, BITS-Pilani, Rajastan, India.

Babu, B.V. (2017). Nature-Inspired Evolutionary Computation for Optimization. i-manager’s Journal on Software Engineering, 12(2), 25-36. https://doi.org/10.26634/jse.12.2.14064

Abstract

As and when the conventional analytical and empirical approaches fail to optimize a given system, we need to look for alternatives. The emerging trend is to get inspiration from nature to handle such complex situations and systems. Traditionally, the gradient based optimization techniques are used for finding an optimal solution to a given problem. However, due to the inability of these techniques in terms of handling multi-variable multi-constrained complex systems, the nature inspired evolutionary algorithms (population based search algorithms) have been developed over the past few years. This paper focuses on nature- or bio-inspired evolutionary computation technique called Differential Evolution (DE) (an improved version of Generic Algorithm), its working principle, and demonstration with a numerical example using step-by-step procedure. Various DE strategies are discussed and applied to many engineering and management problems. DE is extended to multi-objective optimization problems as Multi-Objective Differential Evolution (MODE) and its variants, that can handle the limitations of traditional optimization techniques in addressing complex engineering problems in terms of constraints, objectives, etc. are demonstrated. The working principles of these Evolutionary Algorithms are demonstrated with examples and industrial applications.