A new cache document replacement policy considering the contribution to sales

Most traditional cache document replacement policies are focused on the efficiency aspect and these documents are replaced according to their last access times, request frequencies, and sizes. However, in addition to providing efficient document acquisition service, a successful commercial website also has to create incentives for customers, so as to gain sufficient revenues to support the continuing operation of the website. For this reason, a new cache document replacement policy considering the contribution-to-sales of every document is proposed in this study. In order to evaluate the contribution-to-sales of a document, two web mining techniques are applied. Then the traditional GDSF policy is modified to incorporate this new factor. To evaluate the effectiveness of the proposed methodology, an experimental EC website has been constructed, and a series of computer programs have been developed to simulate the accesses by various kinds of users to the website and to obtain some data for rudimentary analyses. The results showed that the proposed replacement policy could at most increase the hit rate and the byte hit rate by 16% and 9%, respectively, for customers compared with traditional replacement policies.