A Study Of Optimization Methods For Ship Pipe Routing Design And Applications

The ship piping system is as important a system in a ship as a blood vessel system is in a human body. With this system, other various systems in a ship could be connected and function accurately. It is one of the most important design steps in design of ship. In the mean time, ship piping design also influences the economic cost of a ship. Ship piping system design has five consecutive phases: preliminary design, detail design, production engineering, and system-support information. Pipe Routing Design (PRD) is to design the appropriate routes for pipes connecting the start and the goal points in 3D workspace, satisfying various constraints and objectives. Pipe routing is the most important activity during the detail design phase because it takes over 50% of the total detail-design man-hours. At the present time, Ship Pipe Routing Design (SPRD) mainly relies on human experts and their related experience in practice. Due to the great number of ship pipes, highly complex constraints and the large search space, it is very difficult and time-consuming. Compared with the whole ship design in utilizing advanced technology, SPRD is left far behind and become the bottleneck of shortening ship design period. Therefore, any progress toward automating and optimising SPRD will benefit shipyards in terms of higher quality, reduced cost, and shorter turn-around time.Since 1970's, PRD has been a research topic. Starting with the stationary 2D workspace and simple obstacles, research has grown to the stage of 3D dynamic, multi constraint and multi objective problems, resulting in various approaches. These researches are very useful and valuable to the further research of PRD. But until now, there is not a set of mature theory and ideal method yet.Similar to path planning of mobile robot, PRD is a combinatorial optimization problem in 3D space, belonging to NP-hard. Classical optimisation methods usually arouse exponential computational complexities. For such problems the only reasonable way is to look for heuristic algorithms, which find very high quality solutions in short computation time. Genetic Algorithms (GAs) , being one of the metaheuristic algorithms, can be viewed as parallel search technique that simulates the evolution of living things for optimization inspired by natural evolution. Due to its robust, powerfully stochastical and parallel search mechanism, GAs has been applied to many complicated fields of optimizatioin successfully and widely. According to the characteristics of SPRD, this dissertation establishes a Ship Pipe Routing...