Use of a self-adaptive penalty approach for engineering optimization problems
Computers in Industry
Multi-Objective Optimization Using Evolutionary Algorithms
Multi-Objective Optimization Using Evolutionary Algorithms
Some Guidelines for Genetic Algorithms with Penalty Functions
Proceedings of the 3rd International Conference on Genetic Algorithms
A Multi-objective Approach to Constrained Optimisation of Gas Supply Networks: the COMOGA Method
Selected Papers from AISB Workshop on Evolutionary Computing
A flexible tolerance genetic algorithm for optimal problems with nonlinear equality constraints
Advanced Engineering Informatics
A hybrid evolutionary multi-objective and SQP based procedure for constrained optimization
ISICA'07 Proceedings of the 2nd international conference on Advances in computation and intelligence
A penalty function-based differential evolution algorithm for constrained global optimization
Computational Optimization and Applications
| Hi-index | 0.00 |
Equality constraints are difficult to handle by any optimization algorithm, including evolutionary methods. Much of the existing studies have concentrated on handling inequality constraints. Such methods may or may not work well in handling equality constraints. The presence of equality constraints in an optimization problem decreases the feasible region significantly. In this paper, we borrow our existing hybrid evolutionary-cum-classical approach developed for inequality constraints and modify it to be suitable for handling equality constraints. This modified hybrid approach uses an evolutionary multi-objective optimization (EMO) algorithm to find a trade-off frontier in terms of minimizing the objective function and the constraint violation. A suitable penalty parameter is obtained from the frontier and then used to form a penalized objective function. The procedure is repeated after a few generations for the hybrid procedure to adaptively find the constrained minimum. Unlike other equality constraint handling methods, our proposed procedure does not require the equality constraints to be transformed into an inequality constraint. We validate the efficiency of our method on six problems with only equality constraints and two problems with mixed equality and inequality constraints.