| Summary: | 碩士 === 國立臺灣海洋大學 === 系統工程暨造船學系 === 93 === Abstract
The objective of the two-dimensional optimal nesting problem is to determine the effective usage of the stock plate under different considerations of parts. As known, there still exist some artificial approaches for dealing with the nesting problems in the industry such that it usually takes much time or it is quite hard to determine the optimal nesting. Therefore, a set of intelligence-based, automatically nesting system is necessary for the industry.
In this thesis, based on the AutoCAD platform and Genetic Algorithms (GAs) optimizer, a two-dimensional automatically nesting system is proposed to deal with irregular or hollow-empty typed stock plates or parts. Due to the AutoCAD platform existing in this system, an interface with convenient input and interactions for the user is offered and can further be used to the follow up cuttings after the output. As for the GA, it can improve the shortcomings of the artificial selections and further be used to search for the optimal sequence of parts allocations for enhancing the effectiveness of nesting. For the mechanism of this nesting system, it includes three sections, that is, Section 1: the unit resolution is imposed on the configurations of the stock plate and parts such that their configurations can be enclosed by finite strip elements for effectively allocating parts and improving nesting efficiency; Section 2: the adopted Bottom-Left-Fill algorithm, which is developed by improving the Bottom-Left algorithm, is more suitable for the currently developed nesting system; Section 3: the GA is adopted here for the search of optimal sequences of parts. Several testing examples are provided for validating the proposed nesting system and results show that utility rates are all within the admitted errors.
|
|---|
