Combinatorial Coalition Formation for multi-item group-buying with heterogeneous customers
Combinatorial Coalition Formation for multi-item group-buying with heterogeneous customers
复制标题
DOI：
10.1016/j.dss.2009.12.002
复制
发表时间：
2010-04
期刊：
影响因子：
--
通讯作者：
Cuihong Li;K. Sycara;Alan Scheller-Wolf
中科院分区：
文献类型：
--
作者：
Cuihong Li;K. Sycara;Alan Scheller-Wolf
团购市场可能会向物品估价往往存在差异的买家提供具有非附加价值的多种物品（即，物品可能是互补的或可替代的）。在这种情况下，购买群体的形成应集中购买共同商品的购买者，同时考虑购买者对商品捆绑的不同偏好。此外，它应该允许同一团体中的买家之间不统一的成本分摊，这通过吸引更多的团购参与者来使所有买家受益。我们引入了组合联盟（CCF）的概念，它允许买家宣布商品组合的底价。这些底价以及每个商品的卖家价格-数量曲线用于确定每个商品的购买群体的形成。此外，同一群体的买家不一定都支付相同的价格。 CCF 的目标是最大化买家的总剩余。在 CCF 中确定最优联盟配置是 NP 困难的，并且这种配置的稳定性依赖于每个组内的成本分摊规则。因此，我们提出了一种基于增强贪婪选择的 CCF 启发式算法，以及满足某些稳定性属性的成本共享规则。仿真结果表明，与最优结果相比，我们的近似算法生成了相当好的解决方案，并且大大优于更简单的分布式方法。此外，当商品具有互补性或强可替代性时，我们的算法的性能会得到增强，特别是在价格随数量快速或缓慢下降的情况下。 CCF下对卖方收入的评估表明，卖方应为互补性或强替代性商品提供更逐渐下降的价格-数量曲线，为弱替代性商品提供更急剧下降的价格-数量曲线。此外，卖家可能会受益于更简单的价格-数量曲线和更少的步骤所产生的更大销售额。
A group-buying market may offer multiple items with non-additive values (i.e., items may be complementary or substitutable), to buyers who are often heterogeneous in their item valuations. In such a situation, the formation of buying groups should concentrate buyers for common items while taking into consideration buyers' heterogeneous preferences over item bundles. Also, it should permit non-uniform cost sharing among buyers in the same group, which benefits all buyers by drawing more group-buying participants. We introduce the concept of Combinatorial Coalition Formation (CCF), which allows buyers to announce reserve prices for combinations of items. These reserve prices, along with the sellers' price-quantity curves for each item, are used to determine the formation of buying groups for each item. Moreover, buyers in the same group may not necessarily all pay the same price. The objective of CCF is to maximize buyers' total surplus. Determining the optimal coalition configuration in CCF is NP-hard, and the stability of such a configuration relies on the cost sharing rule within each group. We thus propose a heuristic algorithm for CCF based on augmented greedy selections, along with a cost sharing rule satisfying certain stability properties. Simulation results show that our approximate algorithm generates fairly good solutions compared to the optimal results, and is greatly superior to a simpler distributed approach. Furthermore, our algorithm's performance is enhanced when items are complementary or strongly substitutable, especially in settings when the prices decrease either rapidly or slowly with the quantities. Evaluations of the sellers' revenue under CCF demonstrate that sellers should offer a more gradually decreasing price-quantity curve for complementary or strongly substitutable items, and a more abruptly decreasing curve for weakly substitutable items. In addition, sellers may benefit from greater sales generated by simpler price-quantity curves with fewer steps.