Reservoir Operation Optimization using Stochastic Adaptive Refinement of Ant Algorithms

The Algorithm of the Ant Colony Optimisation (ACO) is basically developed and used for discrete optimization problems. However many real engineering problems such as reservoir operation problems are of a continuous nature and using ant based algorithms on such problems requires discretisation of the decision variables. An adaptive refinement mechanism is suggested in this paper to improve the performance of ant algorithms in solving continuous optimization problems. This is an iterative method starting with a uniform discretisation of the search space. A Gaussian distribution is used for discretisation of the decision variables in the subsequent iterations. The average and standard deviation of the Gaussian distribution is computed in each iteration using the optimal solution obtained in the previous iterations. The proposed mechanism was used to solve some benchmark function optimization problems and a reservoir operation problem. The results indicated the efficiency and effectiveness of the proposed method to improve the performance of the ant algorithms for continuous optimization problems.