- 學位論文
基因演算法應用於B2C電子商務配送模型
B2C E-commerce Distribution Model Using Genetic Algorithm
摘要
區域途程問題在生活中無所不在，不同於過去，現在電子商務的形式以更多元方式發生，這也歸功於平台和共享經濟的普及，任何人都可以是平台的供應者和需求者，在平台的幫助下，使得人們更加容易被連結在一起來促進供需雙方的媒合。過去只有企業需要思考如何有效的將產品送達客戶的手中，現在所有人都可以是供給方，路徑問題在日後可以有更廣泛的應用，應該找到一個具時間效率的方式來求得最佳的途程，路徑又可以是各種形式，可以是多重的路徑也可以是對稱或非對稱。 面對多樣的路徑問題，以基因演算法來解決這些問題，只要稍微的修改基因解碼的路徑定義即可以容易的應用在所有途程。基因演算法的運算速度完全取決於是否可以對每個步驟提出有效的演算法，我們也在這篇論文中提到每個細節給予建議，像是動態規劃的應用和參數的設定，基因的生成數與疊代次數則要根據實際碰到的資料而定。從一些實證的資料來看，基因演算法應用在途程問題是相當方便且有許多改進的空間。
並列摘要
The problem of location routing is ubiquitous. Today, the form of e-commerce occurs more diversely. It is due to the popularity of platforms and sharing economy. Anyone can be a supplier and a demander of different open platforms of his choice. Thus, it is easier for people to be linked together to promote the match between supply and demand. In the past, only companies had to consider the question of how to effectively deliver products to customers. However, now anyone can be a supplier, so the problem of routing will be more widely used in the future. It should be essential to use a time-efficient way to search for the optimal route. Routes can take various forms, such as multiple, symmetry, and asymmetry. There are various path problems, and genetic algorithms are applied to solve those problems. Gene decoding can be modified to be easily applied to different forms of routes. The speed of genetic algorithm depends on whether an effective algorithm can be proposed for each step. We also mentioned the details in the algorithm to give advice such as the application of dynamic programming and settings of parameters. The number of initial chromosomes and the number of iterations are still based on the actual data. From some empirical data, the application of genetic algorithms to the problem of routing is quite suitable and there is much room to improve.