Search

Over the last two decades, many Genetic Algorithms have been introduced for solving optimization problems. Due to the variability of the characteristics in different optimization problems, none of these algorithms performs consistently over a range of problems. In this paper, we introduce a GA with a new multi-parent crossover for solving a variety of optimization problems. The proposed algorithm also uses both a randomized operator as mutation and maintains an archive of good solutions. The algorithm has been applied to solve the set of real world problems proposed for the IEEE-CEC2011 evolutionary algorithm competition.

Over the last two decades, many Differential Evolution (DE) strategies have been introduced for solving Optimization Problems. Due to the variability of the characteristics in optimization problems, no single DE algorithm performs consistently over a range of problems. In this paper, for a better coverage of problem characteristics, we introduce a DE algorithm framework that uses multiple search operators in each generation. The appropriate mix of the search operators, for any given problem, is determined adaptively. The proposed algorithm has been applied to solve the set of real world numerical optimization problems introduced for a special session of CEC2011.

Solving a Constrained Optimization Problem (COP) is much more challenging than its unconstrained counterpart. In solving COPs, the feasibility of a solution is a prime condition that requires the conversion of one or more infeasible individuals to feasible individuals. In this paper, to encourage the effective movement of infeasible individuals towards a feasible region, we introduce a Constraint Consensus (CC) method within the Differential Evolution (DE) algorithm for solving COPs. The algorithm has been tested by solving 13 well-known benchmark problems. The experimental results show that the solutions are competitive, if not better, as compared to the state of the art algorithms.

Due to the variability of the characteristics of different Constrained Optimization Problems, no single Differential Evolution strategy, with no single constraint handling technique, performs consistently over a range of problems. In this paper, for a better coverage of the problem characteristics, we introduce a DE algorithm that uses multiple search operators and constraint handling techniques. In the proposed algorithm, initially each individual is assigned a random combination of operators. After a certain number of generations, the improvement made by each combination is recorded, and the best combination is then assigned to more and more individuals, while each of the other individuals are assigned a random combination. To accelerate the convergence of the proposed algorithm, a local search procedure is also applied to selected individuals. The algorithm has been tested by solving 18 test problems, with 10D and 30D. The results showed that the proposed algorithm is superior to state of the art algorithms.

Over the last two decades, many Genetic Algorithms have been introduced for solving Constrained Optimization Problems (COPs). Due to the variability of the characteristics in different COPs, none of these algorithms performs consistently over a range of problems. In this paper, we introduce a Genetic Algorithm with a new multi-parent crossover for solving a variety of COPs. The proposed algorithm also uses a randomized operator instead of mutation and maintains an archive of good solutions. The algorithm has been tested by solving the 36 test instances, introduced in the CEC2010 constrained optimization competition session. The results show that the proposed algorithm performs better than the state-of-the-art algorithms.