Optimization of Multi-Execution Modes and Multi-Resource-Constrained Offshore Equipment Project Scheduling Based on a Hybrid Genetic Algorithm

Abstract

Offshore engineering construction projects are large and complex, having the characteristics of multiple execution modes and multiple resource constraints. Their complex internal scheduling processes can be regarded as resourceconstrained project scheduling problems (RCPSPs). To solve RCPSP problems in offshore engineering construction more rapidly, a hybrid genetic algorithm was established. To solve the defects of genetic algorithms, which easily fall into the local optimal solution, a local search operation was added to a genetic algorithm to defend the offspring after crossover/mutation. Then, an elitist strategy and adaptive operators were adopted to protect the generated optimal solutions, reduce the computation time and avoid premature convergence. A calibrated function method was used to cater to the roulette rules, and appropriate rules for encoding, decoding and crossover/mutation were designed. Finally, a simple network was designed and validated using the case study of a real offshore project. The performance of the genetic algorithm and a simulated annealing algorithm was compared to validate the feasibility and effectiveness of the approach.