Winner(s) determination in combinatorial reverse auctions is a very appealing application in e-commerce but very challenging especially when multiple attributes of multiple instances of items are considered. The difficulty here is to return the optimal solution to this hard optimization problem in a reasonable computation time. In this paper, we make this problem more interesting by considering all-units discounts on attributes and solving it using genetic algorithms. We also consider the availability of instances of items in sellers' stock. In order to evaluate the performance of our proposed method, we conducted several experiments on randomly generated instances. The results clearly demonstrate the efficiency of our method in determining the winner(s) with an optimal procurement cost in an efficient processing time.
展开▼