Binary-coded and real-coded genetic algorithm in pipeline flow optimization

The mathematical model for the liquid-gas mixture flow in pipelines is an initial-boundary value problem for a nonlinear hyperbolic conservation law system. This hyperbolic conservation law system together with boundary conditions is numerically solved using the essentially non-oscillatory (ENO) schemes. The optimization problem is a boundary control problem, i.e. boundary conditions that cause pressure values in the pipeline as close as possible to the desired ones are to be found, considering given constraints. The applied optimization method is the genetic algorithm (GA) with two different variable-to-chromosome coding strategies: binary coding and real coding. The results of both GA strategies applied to two pipeline flow optimization problems are presented and compared.