Application of Adaptive Dynamic Differential Evolution Algorithm in Optimal Allocation of Water Resources (Case Study: Kerman Baft Dam)

Optimal use of dam’s water resources is a serious challenge in water resources management and engineering, due to continuous droughts and water shortages. Therefore, in order to comprehensively management in the used amount of dam’s water resources, it is essential to establish a control mechanism in the release of water resources. In this study, an improved evolutionary algorithm entitled "Adaptive Differential Evolution by Choosing of Dynamic incremental intervals (ADECDII)" was proposed to optimize the single-reservoir water system. Efficiency of the ADECDII approach was to use the dynamic incremental interval selection scheme for parameters adjustment of the classical differential evolution (DE). Problem modeling was defined as a problem of minimization with the objective function of error values between real demand and released water, with one-month time periods between 2008-2018 years, on Baft Dam in Kerman province. ADECDII performance evaluation was compared with six advanced algorithms. Based on the statistical results, lowest average of the dam water shortages with mean error of zero ( ) MCM (Million Cubic Meters) was recorded for ADECDII, which proved the efficiency of the proposed approach in the optimal allocation of released water values. This is while, other comparable algorithms could not report a real shortage of less than 0.37 MCM. The average of runtime for ADECDII was 7.15 sec, which was three seconds longer than DE runtime. Also, ADECDII in the tests of reliability index, vulnerability index, resilience index, sustainability index, difference absolute error values between released water values and total demand in an annual average, and convergence rate clearly showed a higher performance in compared to other comparable algorithms.