An overview of GA and PGA

Abstract

Genetic algorithms have been proven to be both an efficient and effective means of solving certain types of search and optimization problems. Genetic algorithms have been applied with positive results in many areas including scheduling problems, neural networking, face recognition and other NPcomplete problems. The idea behind GA´s is to extract optimization strategies nature uses successfully -known as Darwinian Evolution -and transform them for application in mathematical optimization theory to find the global optimum in a defined phase space. Another popular way to improve genetic algorithms is to run them in parallel, some parallel genetic algorithms have performed very well compared to the standard non-parallel genetic algorithm. Parallel genetic algorithms focus their efforts at simulating multiple species and include not only the standard operations for crossover and mutation but also operations for migration between different populations. Genetic algorithm (GA) which is a meta-heuristic algorithm has been successfully applied to solve the scheduling problem. The fitness evaluation is the most time consuming GA operation for the CPU time, which affects the GA performance. This paper proposes and implements a synchronous master-slave parallelization where the fitness evaluated in parallel. The rest of paper organized as follow: genetic algorithm, parallel genetic algorithm, proposed algorithm, theoretical analysis, practical analysis, and conclusion.

Review on Power Quality Disturbances Detection, Classification, Optimization and Mitigation Techniques

“…GA is an optimization problem-solving meta-heuristic algorithm [106], characterized for the search tasks by imitating genetic behavior and the natural world [107]. The goal of this iterative process is maintaining a population of structures that could potentially solve particular domain challenges .…”

“…The goal of this iterative process is maintaining a population of structures that could potentially solve particular domain challenges . Using particular genetic operators like reproduction, crossover, and mutation [106], a new population of candidate solutions is created during each temporal increment ( generation) based on evaluating the effectiveness of structures in the current population as domain solutions . The population approaches the optimal solution with each increment [R3].…”

Untitled

“…As explained previously, the GA should be implemented in a parallel style to achieve real-time path planning and overcome the GA's computational drawback. Many studies, such as [16][17][18], have discussed the Parallel Genetic Algorithm (PGA). These researches focus on models, hardware architectures, APIs, and middleware used to implement PGAs.…”

Knowledge sharing in cooperative compact genetic algorithm

“…Fitness_Offspring = Fitness_function(Offspring) with each other to discover the high-quality solutions. There are many types of the parallel genetic algorithm such as master-slave, fine-grained and coarse-grained [11], [12]. In the master-slave model, Master has duties to control and share data to all slaves, while slaves have the same duties.…”

“…The problems in the experiment has 3 sizes of cites (11,15, and 17 cities and 5 problems per each). A problem of the 17-cities problems was from TSPLIB, a collection of traveling salesman problem datasets maintained by Gerhard Reinelt [35].…”