Cuevas E., Cortés M.A.D., Navarro D.A.O. Advances of Evolutionary Computation. Methods and Operators

Springer, 2016. — 214.Evolutionary computation (EC) is one of the most important emerging technologies of recent times. Over the last years, there has been exponential growth of research activity in this field. Despite the fact that the concept itself has not been precisely defined, EC has become the standard term that encompasses several stochastic, population-based, and system-inspired approaches.
EC methods use as inspiration our scientific understanding of biological, natural, or social systems, which at some level of abstraction can be represented as optimization processes. They intend to serve as general-purpose easy-to-use optimization techniques capable of reaching globally optimal or at least nearly optimal solutions. In their operation, searcher agents emulate a group of biological or social entities which interact to each other based on specialized operators that model a determined biological or social behavior. These operators are applied to a population (or several subpopulations) of candidate solutions (individuals) that are evaluated with respect to their fitness. Thus, in the evolutionary process, individual positions are successively approximated to the optimal solution of the system to be solved.
Due to their robustness, EC techniques are well-suited options for industrial and real-world tasks. They do not need gradient information, and they can operate on each kind of parameter space (continuous, discrete, combinatorial, or even mixed variants). Essentially, the credibility of evolutionary algorithms relies on their ability to solve difficult, real-world problems with the minimal amount of human effort.
There exist some common features clearly appear in most of the EC approaches, such as the use of diversification to force the exploration of regions of the search space, rarely visited until now, and the use of intensification or exploitation, to investigate thoroughly some promising regions. Another common feature is the use of memory to archive the best solutions encountered.
Numerous books have been published tacking into account any of the most widely known methods, namely simulated annealing, tabu search, evolutionary algorithms, ant colony algorithms, particle swarm optimization, or differential evolution, but attempts to consider the discussion of alternative approaches are scarce.
The excessive publication of developments based on the simple modification of popular EC methods presents an important disadvantage, in that it distracts attention away from other innovative ideas in the field of EC. There exist several alternative EC methods which consider very interesting concepts; however, they seem to have been completely overlooked in favor of the idea of modifying, hybridizing, or restructuring traditional EC approaches.
The goal of this book is to present advances that discuss alternative EC developments or highlight non-conventional operators which prove to be effective in adapting a determined EC method to a specific problem.A Swarm Global Optimization Algorithm Inspired in the Behavior of the Social-Spider
A States of Matter Algorithm for Global Optimization
The Collective Animal Behavior Method
An Evolutionary Computation Algorithm based on the Allostatic Optimization
Optimization Based on the Behavior of Locust Swarms
Reduction of Function Evaluations by using an evolutionary computation algorithm
Collective Animal Behavior Algorithm for Multimodal Optimization Functions
Social-Spider Algorithm for Constrained Optimization