Comparison of optimization methods for irrigation systems based on suboptimal solutions

This paper deals with optimizing sprinkler irrigation systems, especially their most expensive part, i.e., the water distribution pipeline network. The authors note that the optimized design of a pipe network is usually accomplished by several runs of some heuristic evolutionary algorithm. This process takes up quite a lot of time. Such a series of calculations aims to find the heuristic algorithm's most suitable (problem depending) parameters for the final optimization. However, this procedure does not guarantee finding the optimal solution, even after many runs of algorithm. The authors propose and compare four methods that use suboptimal solutions found in the trial-and-error phase of the mentioned optimization process that speed up the search for a final solution and helps find a solution with a lower cost than the solutions achieved if continuing with this standard process. The proposed calculation procedures were tested and compared on a large irrigation system.