由於網路資訊透明度高且容易取得,電子商務環境也日益成熟,造就了許多不同型態的網路交易模式的興起,網路上的雙向拍賣便是其中之一。雙向拍賣交易機制由買賣雙方分別對產品出價,透過市場的運作機制協助買賣雙方進行撮合,達到最後成交的目的。網路上的雙向拍賣交易研究主要是藉由智慧代理人技術來達成,包括市場機制的設計、買賣雙方都是由智慧代理人來操控,本研究將以交易代理人的觀點,提出在不同的市場環境下皆能自我調適的交易代理人機制,並透過模擬的方式來驗證其運作效率。本研究之交易代理人程式以基因表達編程法(GEP)為核心,透過其演算法的自我適應特色達到交易代理人可隨著不同市場環境選擇適當交易策略的目的,研究實驗在TAC競賽(Trading Agent Competition)所提供的模擬平台進行,結果發現本研究所提出的交易代理人能夠維持交易績效的穩定度,並在買方市場中具有良好的交易績效。
With the information transparency and accessible in the internet, more and more effective business models of electronic commerce (EC) have been proposed. Double auctions transaction model in EC has been proposed and attracts much attentions. In the double auctions market, the traders, including sellers and buyers, can quote by themselves, and the double auctions market matches the sellers and buyers by designed transaction rules automatically. In this research, we design an intelligent agent for the traders based on Gene Expression Programming (GEP) to transact in the market robotically. The GEP based traders can adapt their transaction policies to fit the inconstant market. Our experiment has been conducted in Trading Agent Competition (TAC) platform. The results show that our traders can gain more stable transaction performance in the double auctions market.
為了持續優化網站功能與使用者體驗,本網站將Cookies分析技術用於網站營運、分析和個人化服務之目的。
若您繼續瀏覽本網站,即表示您同意本網站使用Cookies。