Bilevel optimization using bacteria foraging optimization algorithm

Abstract

Bilevel programming problems involve two optimization problems where the constraint region of the first level problem is implicitly determined by another optimization problem. There are number of different algorithms developed based on classical deterministic optimization methods for Bilevel Optimizations Problems (BLOP), but these are very much problem specific, non-robust and computation intensive when number of decision variables increase, while not applicable for multi-modal problems. Evolutionary Algorithms are inherently parallel, capable of local as well as global search, random, and robust techniques and can used to solve these BLOPs. In this paper, Bilevel Bacteria Foraging Optimization Algorithm (BiBFOA) is proposed for solving BLOP based on the foraging technique of common bacteria. Experimental results demonstrate the validity of the BFOA-based algorithm for solution of BLOPs.