Analysis of Nonlinear Optimization Problems Using Differential Evolution Algorithm

Differential evolution algorithm (DE) is a stochastic, population-based optimization approach for resolving nonlinear optimization issues. Differential evolution approach (DE) is a straight hunt technique parallel in nature with three decision parameters, namely, crossover rate, scaling factor, and population size. It is the progression of generation phase, mutation, and crossover followed by selection. DE has been useful to solve large number of engineering design issues together as single as well as multi-objective optimization approaches. DE implements a greedy assortment method. The superior one of the innovative outcome and the parent of aforementioned succeed the rivalry attaining best converging performance over the other evolutionary computation methods. DE algorithm is a speculative optimization strategy that can model the problem objectives while including constraints to minimize the objective function. The unique features of this approach are obtaining the accurate global minimum independent of the preliminary constraint values, hasty convergence characteristics, utilizing lesser number of control variables, flexible for discrete and integer optimization, efficient in solving the nonlinear constraint optimization like consequence functions, as well as multimodal hunt spaces. In the searching process, the population of DE moves through various regions in the search space, during which certain schemes allied with the specific variable settings are added efficiently than the other schemes. As a result, it is advantageous to flexibly establish a suitable strategy and its related variables at various phases of the advancement procedure. This chapter concentrates on a category of metaheuristic technique called the differential evolution approach to work out valid parameter optimization challenges. New techniques and algorithms to improve the standard DE algorithm to deal with various complex problems are investigated. Classification-assisted DE by incorporating classification rather than ranking or regression for pairwise comparison into DE for solving high complex functions is discussed. Various methods to boost the concert of the conventional DE algorithm like domineering the population diversity of DE through inserting new variables to calculate the assortment all through the progression practice are studied. For certain applications relating single and multi-objective optimization model based on this algorithm is discussed. The efficacy of the variation system for the parameters utilizing yardstick techniques to enhance the overall concert compared to other methods is studied here. Up to date, enhancements in this algorithm along through case studies with real-time application are also analyzed in this chapter. Thus, a thorough study of the differential evolution approach for solving the nonlinear optimization challenges is elaborated in this chapter.