用改进遗传算法求解组合拍卖竞胜标

从电子商务中的组合拍卖机理出发,以第一价格密封拍卖方式为背景,通过分析组合拍卖标的集和竞胜标确定的复杂性,给出了组合拍卖竞胜标确定问题的一般模型,并指出了该问题为离散组合优化问题·同时针对拍卖实践中组合标出现的事实,对求解该模型的单亲遗传算法的初始种群进行优化设计,使得可行解的搜索空间大大缩小·基于这种思想,提出了一种适合求解该模型的改进遗传算法·计算实例表明,利用该算法求解竞胜标确定问题的最优解,算法具有实现简单、寻优速度快、计算效果良好等特点·

Improved Genetic Algorithm for Solving Winner Determination in Combinator ial Auctions

Starting from the mechanism of combinatorial auction in electronic commerce, the complexities of both the set of target and winner determination were analyzed against the background of the sealed first-price auction, so as to give a general model for winner determination in combinatorial auctions. The model presented itself to be a problem of discrete combinatorial optimization. Optimizing the initial population of the partheno-genetic algorithm to solve this model, the search space of feasible solutions was greatly shrunk so as to adapt to the combinatorial targets coming out in auctions. An improved genetic algorithm for solving this model was thus presented on the basis of the idea as above, which is simple to apply and can achieve rapidly fine outcome, showing batter computation result than the previous ones.