Two-stage genetic algorithm for parallel machines scheduling problem: Cyclic steam stimulation of high viscosity oil reservoirs

• Published in: Applied soft computing 2018-03, Vol.64, p.317-330
• Main Authors: Sheremetov, Leonid, Martínez-Muñoz, Jorge, Chi-Chim, Manuel
• Format: Article
• Language: English
• Subjects: Cyclic steam stimulation; Genetic algorithm; Oilfield; Scheduling problem
• ISSN: 1568-4946; 1872-9681
• DOI: 10.1016/j.asoc.2017.12.021

[Display omitted] •A novel practical application of the classical Parallel Machines Scheduling (PMS) problem with release dates in petroleum engineering is considered.•A two-stage genetic algorithm is proposed for solving the PMS problem.•The use of heuristic algorithm for the generation of initial populations along with the design of the chromosome improve the convergence.•The design of the chromosome enables working only with feasible injection schedules.•The proposed approach is evaluated using a real-world data set from the oilfield asset located in the coastal swamps of the Gulf of Mexico. In this paper, the problem of optimal assignment of trailer-mounted steam generators for cyclic steam stimulation (CSS) of petroleum wells is formulated as a parallel uniform machines scheduling (PMS) problem with release dates. The total weighed tardiness is used as the goal of the optimization process. The distinctive features of the proposed PMS formulation include: jobs with variable weights, variable machine setup time, constraint capacity of drilling pads (where machines are allocated), and modified tardiness criterion. A two-stage scheduling algorithm combining heuristic and genetic algorithms is proposed for solving it. A chromosome representation, crossover and mutation operators generating only feasible solutions and thus avoiding the use of any repair mechanism are discussed. The performance of the algorithm is tested on a real-world data set from the oilfield asset located in the coastal swamps of the Gulf of Mexico. The experiments indicate that the proposed approach gives good results in optimization of the operational costs and petroleum recovery. The algorithm is implemented as a part of the software platform for optimization of CSS and currently is in use by oilfield engineers.