Résumé
KDW-GA is a framework for the simulation of preferential water flow through unsaturated soils. Preferential flow can be described by a Kinematic Dispersive Wave (KDW) equation which depends on three transport parameters. Transport parameters are estimated with the binary genetic algorithm (GA) inverse method by reducing the errors (cost function) between estimated and observed water flux values. Different GA components are discussed in order to find the best strategy that fits our problem. Recommendations concerning the mutation rate, the elite number, and the pairing technique are deduced with regard to the algorithm performance.