Model and Algorithm of Winner Determination Problem in Multi-item E-Procurement with Variable Quantities

Reverse E-auction has been widely applied to the centralized E-procurement of governments and large enterprise groups recently. In this paper, a multi-item procurement problem with variable quantities is investigated and the organization process of the corresponding reverse E-auction is introduced. The winner determination problem (WDP), which plays an important role in the investigated problem, is studied in detail. To address this WDP, a mathematical model, which aims to maximize the expected economization of procurement, is established and a solution algorithm based on genetic algorithm (GA), where an order encoding scheme is designed and a special repair method is employed to accomplish the translation from the individual encoding to the corresponding solution of WDP, is proposed. On a test case of WDP which is constructed randomly, experiments are carried out to examine the validity of investigated model and algorithm.