An Intelligent Invasive Weed Optimization: a Q-learning Approach

Invasive Weed Optimization is a recently proposed population based meta-heuristic that mimics the colonizing action of weeds. In this article, an improvement to the classical algorithm has been proposed by introducing a constriction factor in the seed dispersal stage. Temporal Difference Q-Learning has been employed to adapt this parameter for different population members through the successive generations. The proposed memetic approach, named Intelligent Invasive Weed Optimization (IIWO) has been tested extensively on a set of 15 benchmark functions as well as the real world Circular Antenna Array Design problem. The results indicate the efficacy of our proposed approach.