Comparison of design optimization algorithms of a multiply fractured horizontal well

The paper is devoted to comparison of multiple-objectives optimization algorithms in application to the problem of design optimization of a multiply fractured horizontal well (MFHW). The problem is stated either as a single-objective one, where only the income based on Net Present Value (NPV) is maximized, or as a multi-objective problem, where it is necessary to simultaneously find extremes of NPV, the post-fracture oil production and fracturing costs. Three popular stochastic optimization methods are considered: genetic algorithms (GA), simulated annealing (SA) and particle swarm optimization (PSO). Since PSO, SA and GA techniques employ different strategies and computational efforts, the comparison of their efficiency was carried out by testing on synthetic problems and then applied to the example of a MFHW in a low-permeable oil reservoir.