Colony-mutated ant system for pipe network optimization

A new ant algorithm, namely Colony-Mutated Ant System (CMAS), circumventing the premature convergence phenomenon is proposed in this paper and applied to pipe network optimization problems. The method uses a simple but effective mechanism, namely Pheromone Replacement Mechanism (PRM), to make sure that the global-best solution path always has the maximum trail intensity. This mechanism introduces enough exploitation into the method and, more importantly, enables one to exactly predict the number of global-best solutions at each iteration of the algorithm without the necessity of calculating the cost of the solutions created. This number is used as a measure for premature convergence of the method at each iteration. The colony is then mutated such that a predefined number of global-best solutions survive the mutation process. Two different mutation mechanisms, namely one-bit and uniform mutation are introduced and used. The probability of mutation is adjusted at each iteration so that the required number of global-best solutions survive the mutation. The method is shown to produce results comparable to Max-Min ant system (MMAS) algorithm, while requiring less free parameter tuning. The application of the method to a benchmark example in the pipe network optimization discipline is presented and the results are compared. The results indicate that the proposed CMAS method shows improved performance with improved convergence characteristics. Furthermore, the method requires less computational effort for tuning purposes due to the fewer number of free parameters compared to the MMAS method