Title: Constrained particle swarm optimisation for sequential quadratic programming

Authors: Zach D. Richards

Addresses: United Launch Alliance, Littleton, Colorado 80127, USA

Abstract: Sequential quadratic programming (SQP) methods are commonly used to solve constrained non-linear optimisation problems. However, in recent years there has been great improvement in using evolutionary algorithms to solve non-linear optimisations problems. The difficulty has been determining a correct method for implementing evolutionary algorithm for a non-linear optimisation problem with constraints. In this paper, we are combining the strengths of the traditional SQP method with an evolutionary algorithm, particle swarm optimisation (PSO) for solving a constrained non-linear optimisation problem with equality and inequality constraints. We propose a constrained PSO algorithm be used to solve the quadratic programming (QP) subproblem within the SQP method.

Keywords: nonlinear optimisation; particle swarm optimisation; sequential quadratic programming; SQP; infeasibility function; quadratic subproblem; hybrid optimisation; nonlinear constrained PSO; NLCPSOA.

DOI: 10.1504/IJMIC.2009.030083

International Journal of Modelling, Identification and Control, 2009 Vol.8 No.4, pp.361 - 367

Available online: 09 Dec 2009 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article