Formulation of a nonlinear water distribution problem with water rationing: an optimisation approach

Abstract

A mixed-integer nonlinear programming water distribution problem that incorporates water rationing is presented. The city of Bulawayo water distribution problem is implemented and solved using max–min ant system, genetic, tabu search, and simulated annealing algorithms with 100 runs performed for each algorithm. The results show that the city of Bulawayo can save $3158 a day. The max–min ant system produced the best optimal costs compared with the other algorithms. The least run time is obtained by implementing the tabu search algorithm. Water lost through hoarding during water-rationing periods contributes significantly to the total operational costs. Statistical analysis of the results obtained by different algorithms shows that the optimal costs obtained by tabu search, and simulated annealing algorithms are insignificantly different. Future research may be directed toward incorporating priority among water users and formulating a hybrid algorithm that uses both the max–min ant system and tabu search algorithms to solve such problems.