A Study on the Optimization of the Hull-block Construction Process by Multi-Objective Genetic Algorithm

Abstract

In modern shipyards, the block construction method is widely used in building ship hulls. But with this method, when work falls behind schedule the results include downtime, cost increases, and delivery delays. Therefore, an efficient schedule is strongly required. However, a large number of blocks with different sizes and shapes and different procedures must be considered. In addition, the existence of several objectives, such as total cost reduction, shorter completion period, efficient workforce allocation, make the problem more difficult. The block construction schedule is, in short, so complex that veteran engineers usually resort to a rule of thumb in preparing it. For these reasons, an obtained schedule is not necessarily an efficient one. To improve this situation, we must optimize schedules rationally and obtain practical solutions that satisfy all of these requirements simultaneously. Thus, the Multi-Objective Genetic Algorithm (MOGA) is utilized in this study. MOGA applies a Genetic Algorithm (GA) using the concept of Pareto optimization for multi-objective optimization. This method uses the GA characteristics of collective evolution of a solution group. This makes it possible to efficiently arrive at a practical solution from an almost infinite number of solution candidates. This study produced useful information, such as the trade-off relation between overtime work and a shortened completion period, the Pareto solution for total workforce and total completion period, and optimum worker assignment.