「銷售者」與「夜市」的商業環境一直和人們的生活息息相關,由於此種商 業行為提供了人們低價消費與生活上面的便利,因此使得銷售者與夜市能夠在各 個都市之中蓬勃的發展,也影響了整個都市的發展與計畫。 現今,台灣各城市及鄉鎮區,幾乎都有一個或多個屬於當地人民夜間休閒活 動的市集地,也因為這樣,目前已經有許多業者投入這行業,勢必將導致銷售者 間的過度競爭,進而嚴重衝擊銷售者,導致業者營業額一年不如一年。因此現今 銷售者為了提高利潤而必須改變產品的差異性、增加產品販售的種類數量、增設 擺攤的地點、增加到外縣市設攤機會、注意產品的特性等方法,以增加其利潤。 本論文研究週期性市集問題,將以人工智慧法,包括:免疫演算法、基因演 算法,研究銷售者如何選址市集地點,在不同之民眾產品需求週期、購買距離指 數(影響鄰近民眾前往該市集地點之意願的高或低)下,使探討的總銷售時段所賺 取之收益為最大。
Business environment between “trader” and “night market” is constantly and closely related to people’s lives because such a commercial activity provides people low-price and convenient consumptions. Thus the vigorous development of “trader” and “night market” will lead to the efficiency of development and planning for cities.. Today, for each city and township area in Taiwan, there is one or more night markets for residents. Therefore, there are many traders devoted to such night market activities. However, more traders will result in lower profit. Thus, to achieve higher profit, each trader has to adopt more strategies to earn more profits, including (i) making more difference among their products and the other products, (ii) selling more types of products, and (iii) selling products in more various cities. This thesis studies the periodic market problem. We will apply various artificial approaches, including immune algorithm and genetic algorithm, for solving the periodic market problem. The purpose of the periodic market problem is to find the locations of cities for a trader during a specific time period such that the profit is maximized.