Page 341 - Decision Making Applications in Modern Power Systems

Heuristic methods for the evaluation of environmental impacts Chapter | 12 303

The heuristic optimization techniques can be of exhaustive and nonex-
haustive types. The comprehensive or exhaustive techniques, such as algo-
rithmic schemes, Backtracking and Branch & Bound, have the advantage of
finding the optimal solution always, using the worst case—the entire solution
space is huge. It is difficult to narrow the search by the use of heuristic tech-
niques and, therefore, may result in inefficient algorithms for medium-to-
large problems.
The nonexhaustive techniques are known by the name of metaheuristics,
which can be algorithmic schemes based on different ideas in many outlets,
occasions, and the workings of nature, which is a common approach of
problem-solving by successive improvements of a solution or set of solu-
tions, with an exploration of broader solution space and with some random
factor [6,7].
In this work the metaheuristic techniques, specifically GAs, will be used.
It is taken into consideration that the types of optimization problems have a
very complex resolution space; therefore exploring it completely may not be
feasible for certain applications. In this type of technique, what is done is to
work with a solution or a set of solutions for new responses that are closer to
the optimal in order to avoid the great places and, iteratively, to achieve a
high-quality convergence. In this way, it is possible to guarantee the quality
of the solution, as this will comply with the criteria found.

12.2.2 Genetic algorithms

GAs are adaptive heuristic search algorithms that are based on evolutionary
ideas of natural selection and genetics. As such, they represent an intelligent
exploitation of a random search used to solve optimization problems.
Although randomized, GAs are not random, instead, exploit historical infor-
mation to direct the search for the best performance region within the search
space. The basic techniques of GAs are designed to simulate processes in
natural system necessary for evolution, especially those that follow the prin-
ciples established by Charles Darwin first—“survival of the fittest,” where,
in nature, in competition among individuals for scarce resources, the more
capable individuals dominate over the weak.
It is better than conventional techniques of artificial intelligence (AI) that
is more robust. Unlike older systems, AI, they don’t break easily even if the
inputs change slightly, or in the presence of reasonable noise. In addition,
when searching a large state space, multimodal state space, or n-dimensional
surface, a GA can provide significant benefits on the types of most typical
search engine optimization techniques (linear programming, heuristic depth-
first search width, and praxis) [9].
GAs mimic the survival-of-the-fittest individuals from every successive
generation of a problem to solve. Each generation consists of a population of
strings of characters that are similar to chromosome that we see in DNA.