| 摘要: | 隨著社會經濟的發展，民眾的生活型態有著逐漸的改善，這也使得休閒旅遊已成為民眾重視的生活議題。顧客關係管理(Customer Relationship Management, CRM)是一項近幾年最受重視的研究領域，要如何挖掘潛在顧客與如何快速地開發新顧客通常是飯店業者最積極投入資源進行探討，飯店業者可以根據過去顧客的交易記錄進行有意義的規則或是模型等資料探勘(Data Mining, DM)，再根據所挖掘的結果進行最適行銷建議。人工智慧技術(Artificial Intelligence, AI)在近幾年在許多領域（例如：預測、辨識、控制、分類等）都有不錯的應用成效，其中的類神經網路(Artificial Neural Networks, ANNs)方法更是受到多數研究者的青睞，尤其是資料挖掘中的聚類分析(Clustering Analysis)，類神經網路也因為它的便利性和簡易性而有不錯的成效，在本論文中亦將引用類神經網路中非監督式的自適應性共振理論模式(Adaptive Resonance Theory, ART)來進行必要的顧客屬性群的聚類分析。此外，我們也設計一個指標與評比程序來找出到底哪些舊顧客屬性群對飯店業者而言是最適合進行資源投資以及行銷開發、並利用舊顧客的屬性區隔資訊來對新顧客提供最適行銷建議。 此外，飯店服務人員可能會被詢問到顧客想去玩的旅遊景點應如何搭車較方便？怎麼進行是最省錢？這些顧客服務就要考慮到前往各個景點移動時間或者移動費用（成本）的最適性問題，本研究將應用基因遺傳演算法(Genetic Algorithm, GA)來規劃出旅遊行程最適解，求出顧客欲前往的各個景點總移動時間最少的最佳路徑、總移動成本最少的最佳路徑、以及總移動時間與總移動成本同時為最少的最佳組合路徑，提供顧客旅遊行程規劃參考，以提升飯店業者在市場的競爭力。
As for the social economic rapidly developing, the living style of people in Taiwan had been improved. It will let most people pay more attention to the leisure tour. Recently, customer relationship management (CRM) is a new issue to be addresses for many Hotels or Tourism centers. How to efficiently mine the potential customers or attract the new customers are two primary goals to invest in studying for them. The marketing suggestions can be made by mining the meaningful rules or models from the historical transaction records. Up to now, the performance of using the techniques of artificial intelligence (AI) had presented well for many applications, e.g. prediction, recognition, control, classification and clustering. Among them, the issue of clustering analysis can be done well by using the artificial neural networks (ANNs). Especially, the adaptive resonance theory (ART) in the unsupervised model is the popular one with the characteristics of simplification to be applied into clustering analysis. Hence, we will employ it to implement the necessary clustering analysis in this study. Besides, an evaluated index and a complete evaluation procedure are developed to obtain the related conclusion from the results of DM. That is, the information about the clusters of the potential customers being suitable to invest and that about the corresponding cluster being suitable to develop can be obtained. Besides, the attendant will be frequently asked about how to travel conveniently by public transits? or how to travel with the less cost? Such cases will focus on the problems of the adaptive tour scheduling with moving time or moving cost limitations. Genetic algorithm (GA) is an approach to address the optimization problem well in most industries. In this thesis, we will apply GA to address such problem. The suggestions of adaptive tour scheduling with the minimum tour’s moving time, minimum tour’s moving cost or simultaneously minimize the tour’s moving time and moving cost can be derived. Not only the customer’s requirement can be achieved, but the necessary market strategies can also be created in the subsequent customer relationship management (CRM). The service competence for hotel can be then promoted. |
