Development of a Combined Coding Scheme for a Hybrid Genetic Algorithmwith Sequential Quadratic Programming (GA-SQP) and a Bat Algorithm (BA) for Mixed Integer Nonlinear Programming (MINLP)
International Journal of Advanced Research in Engineering and Technology (IJARET), 11(3), 2020, pp 400-409.
10 Pages Posted: 28 Apr 2020
Date Written: April 3, 2020
This study seeks to optimize the process of mixed integer nonlinear programming (MINLP) using BA_GA-SQP, which is a coding combination of a hybrid genetic algorithm with a bat algorithm (BA), using sequential quadratic programming (GASQP). Initially, the population is divided into discrete and continuous variables, whereupon optimization of the discrete variables is performed using the adapted BA, while the GA-SQP is employed to optimize the continuous variables, making use of discrete variable information obtained from the adapted BA. This allows a smaller population to be set than would usually be the case in order to reach a global solution. A number of common MINLP problems from the field of chemical engineering were used to verify the proposed BA_GA–SQP algorithm. In the problems tested, the fitness values were able to achieve the global optimum.
Keywords: Bat algorithm, Genetic algorithm, Mixed integer nonlinear programming, Sequential quadratic programming, Optimization
Suggested Citation: Suggested Citation