An Improved Fitness Based Differential Evolution Algorithm (IFBDEA)
Hitesh Sharma, Vikas Sharma, Akrati Sharma, Prakash Meena · 2016
Research workers solve the simple problems of optimization by using various mathematical techniques. But to solve complex problems of optimization a stochastic and population based algorithm named Differential Evolution (DE) is used. DE is fast, simple and straightforward algorithm to optimize the problems which are complex. Like other evolutionary algorithms DE also has some drawbacks. In FBDE author tried to overcome these drawbacks and tested on some benchmark functions and real world problems. FBDE is also found good as compared to DE and its recent variants. The proposed algorithm is a improvement of FBDE using my previously published algorithm PGLFDE. In PGLFDE to make the balance between DE's exploration (globally) and exploitation (locally) capabilities levy flight search is modified using different mutation strategies for different population. In proposed algorithm we are hybridizing the PGLFDE into the FBDE to improve the accuracy, efficiency and reliability of FBDE and its name is called IFBDEA. We have plotted the graph between different points of crossover rates of IFBDEA and the sum of success rates of the problems taken. The proposed algorithm has been tested on 23 standard benchmark problems which includes real world problems and personified with DE, LFDE and FBDE algorithms.