Groundwater Management in Coastal Regions Using Linear Programming and Genetic Algorithms

In the past optimization techniques have been combined with simulation models to determine cost-effective solutions for various environmental management problems. In the present study, a groundwater management problem in a coastal karstic region in Crere, Greece has been studied using classical (linear programming) and heuristic optimization methods. First, a numerical simulation model of the unconfined coastal aquifer has been developed to represent the complex physical system. Second, the classical linear programming optimization algorithm of the Simplex method is used to solve the groundwater management problem where the main objective is to control the saltwater intrusion hydraulically. A piecewise linearization of the non-linear optimization problem is obtained by sequential implementation of the Simplex algorithm and a convergence to the optimal solution is achieved. Third, the solution of the non-linear management problem is also obtained using a genetic algorithm. A differential evolutionary algorithm that emulates some of the principles of evolution is used. Finally, A comparison of the results obtained by the two different optimization approaches is performed. A sensitivity analysis is also employed in order to examine the influence of the active pumping wells in the evolution of the seawater intrusion front along the coastline.