Optimal design of D-town network using multi-objective harmony search algorithm

For optimal design of water distribution system in the D-Town, harmony search algorithm, one of meta-heuristic methods, is applied as a methodology to examine the relationships between the design parameters such as cost, greenhouse gas emissions, and water age minimization, and to satisfy the given objective function. The objective function was decided to draw a conclusion that evenly satisfies three design parameters, since the trade-off relationship between the parameters was not strong. So in order to achieve water age minimization, the size of tank was first minimized and then sizes of the pumps and pipes were decided to minimize the sum of cost and greenhouse gas emissions. The skeletonized pipes were used to decide the size of pumps for increasing efficiency of the optimization, and pipe size in each sub-network was optimized to reconstruct the network in each section, in which the decided pump size was implemented. The pipe size was decided by optimization in each section, and considered as initial solution set, to perform the pipe size optimization on the whole network. With the finalized pipe network, the size of diesel generator was decided to satisfy the given constraints for pump flow check, control modification process, and power outage simulation. As a result, the minimum cost, the minimum GHG emission, and the minimum water age, all of which satisfies the given hydraulic constraints for the optimized network, turned out to be 928,951dollars, 2,600,656 kg - CO_2-e, and 0.193hrs, respectively. When evaluated for the appropriateness, such results with subdivided network seemed to draw better solutions than did the results when every pipe in the network was optimized at once.