A multi-strategy multi-population cooperative optimization algorithm for drinking water pollution source identification

Abstract

The occurrence of sudden drinking water pollution incidents can cause severe disasters and significant social losses. To ensure real-time monitoring of water quality and safeguard drinking water safety, it is essential to promptly detect pollution, locate the pollution source, and implement emergency responses. The accuracy of pollution source identification directly determines whether safety risks can be minimized. Existing optimization algorithms for solving the water pollution source identification problem often struggle to balance convergence and diversity and are trap into local optima. As a result, their performance is often unstable. To overcome these defects, this paper proposes a multi-strategy multi-population cooperative optimization algorithm (MMCOA) for drinking water pollution source identification. Based on the characteristics of the drinking water pollution source identification problem, multiple improved search strategies are designed to guide the algorithm in global search. Experimental results show that, under various pollution scenarios in pipeline networks of specific scales, the proposed strategies are effective, and the accuracy of source identification achieved by the proposed algorithm significantly surpasses that of other comparable algorithms.