No abstract

A robust variant of invasive weed optimisation (IWO) algorithm, called enhanced invasive weed optimisation (EIWO) algorithm, is proposed in this paper for the optimisation of constrained benchmark problems. Enjoying the ecological behaviour of colonising weeds, IWO has demonstrated its ability in solving different optimisation problems. Since making a proper balance between these two components is essential, especially to cope with constraint optimisation problems, two new rules are added to the algorithm to improve its performance. The first rule is utilising principles of social standard deviation as proposed in social harmony search (SHS) algorithm. The second rule is utilised to prevent the algorithm to get stuck on local optima. Finally, for constraint handling, three simple heuristic rules of Deb are utilised. The robustness and effectiveness of the proposed method are tested on many constrained benchmark problems and compared against those of state-of-the-art algorithms.Reference to this paper should be made as follows: Ramezani, P., Ahangaran, M. and Yang, X-S. (2013) 'Constrained optimisation and robust function optimisation with EIWO', Int. J.

In this paper, after a literature review, studies will be concentrated on standard deviation of invasive weed optimization's normal distribution function which is used for distributing seeds of each weed over the search space. Although invasive weed optimization is a great algorithm to solve real world practical optimization problems but there is a serious drawback in distributing the seeds over the search space. A new concept will be presented to distribute seeds of each weed over the search space which increases the robustness and effectiveness of algorithm, and therefore leads to an improved invasive weed optimization. Simulation on a set of unconstrained benchmark functions reveals the superiority of the proposed algorithm in quick convergence and finding better solutions compared to the original invasive weed optimization.