Bio-inspired Computation Algorithms

Abstract

Bio-inspired computation is the use of computers to model the living phenomena and simultaneously the study of life to improve the usage of computers. Swarm behaviors in animal groups such as bird flocks, bees, ants, fish schools, and sheep herds, as well as insects like mosquitoes, ants, and bees, often exhibit incredible abilities to solve complex problems that seem far beyond their capabilities. This chapter mainly focuses on the biological inspiration, principle, and implementation procedures of four popular bio-inspired computation algorithms including ant colony optimization (ACO), particle swarm optimization (PSO), artificial bee colony (ABC), and differential evolution (DE). Special emphasis has been laid on how the biological behavior can be transferred into a technical algorithm. Moreover, description of algorithms in more general terms and the most successful variants of these algorithms are provided. Finally, a brief introduction to other bio-inspired computation algorithms such as glowworm swarm optimization (GSM), bacteria foraging optimization (BFO), bat-inspired algorithm (BA) is presented. IntroductionBio-inspired computation, short for biologically inspired computation, is a way of developing computer systems by taking ideas from the biological world. It is a field of study that loosely knits together subfields related to the topics of connectionism, social behavior, and emergence. Briefly put, it is the use of computers to model the living phenomena and simultaneously the study of life to improve the usage of computers. Bio-inspired computation in an interdiscipline composed of many different fields, such as biology, computer science, physics, mathematics, and genetics. As we have known, biological systems have many advantages over computer systems, as they are able to solve much more complex problems beyond H. Duan and P. Li, Bio-inspired Computation in Unmanned Aerial Vehicles,

Autonomous Multi-Drone Warehousing through Deep Reinforcement Learning and Predictive Potential Fields

“…Concerning the reward function shown in Table 1, at every time step of ∆ = 0.2 seconds, the learning agent receives a time penalization = −∆ and a positive reinforcement proportional to the subgoal distance to encourage an efficient task completion. We limit the distance to subgoal to 0.75 in (7) to avoid an excessive positive reward near the target, preventing the agent from completing the subtask. Note that = 1 herein.…”

“…In the last decade, decentralized MAS have been incorporating a large spectrum of techniques, including bioinspired [5][6][7][8][9] , theoretic games [10][11][12][13][14] , and learning based approaches [15][16][17][18] . While the former methodologies can be robust and computationally unexpensive, agents and resources are often inefficiently allocated.…”

Adaptive task allocation for multi-UAV systems based on bacteria foraging behaviour

“…Bio-inspired algorithms are based on the natural behaviour of simple agents as they interact amongst themselves and exhibit favourable features, such as self organisation, adaptiveness, and robustness [25]. This natural behaviour is governed by simple rules that support their implementation and economical execution.…”