This chapter addresses nature and bio-inspired metaheuristics in the context of conflict detection and resolution problems. An approach is presented for a generalization of a population-based bio-inspired search and optimization algorithm, which is depicted for three of the most well-known and firmly established methods: the genetic algorithm, the particle swarm optimization algorithm and the differential evolution algorithm. This integrated approach to a basic general population-based bio-inspired algorithm is presented for single-objective optimization, multi-objective optimization and many-objective optimization. A revision of these three main bio-inspired algorithms is presented for conflict resolution problems in diverse application areas. A bridge between feedback controller design, genetic algorithm, particle swarm optimization and differential evolution is established using a conflict resolution approach. Finally, some perspectives concerning future trends of more recent bio-inspired meta-heuristics is presented.
TopNature And Bio-Inspired Search And Optimization Techniques: Key Issues
This section, begins by overviewing three of the most successful bio-inspired search and optimization techniques: genetic algorithm (GA), particle swarm optimization (PSO) and differential evolution (DE). The methodology used to present key issues concerning these algorithms is based in an integrated approach, by focusing firstly in the common issues to all three algorithms and then in the particular differences regarding each meta-heuristic. Due to the huge number of existing variants and refined versions presented in the last decades for all three algorithms, including hybridization techniques, a simplified approach presenting the core of these bio-inspired algorithms is presented here, in order to make it easier their application to solve conflict resolution problems.
Most applications regarding conflict resolution or any other type of problems, require solving an optimization problem with one or more functions, which can be formulated considering the minimization case as follows:
(1) with:
x(t)=(x1(t),x2(t), …,xn(t)) representing a
n-dimensional decision variable and a potential solution for the problem,
t the evolutionary iteration,
S the feasible search space, g
i the
ith constraint and
fo an objective function. The number of objective functions is denoted as
o, and the number of constraints is denoted as
k (for simplicity sake, equality constraints are not represented in (1)). Three of the most relevant aspects, which influence the most the complexity of the search procedure involved in an optimization problem are: