Reource Allocation Among Multiple Selfish Agent

It is very important for both economics and computer science to effectively and fairly allocate resources among multiple self-interested agents based on their prefer-ences on these resources. In computer science and artificial intelligence, besides econ-omy efficiency and social equality, we are also very interested in "computational prob-lems", i.e. problems of designing effective algorithms to get a "good" allocation, and some other related computational problems. The paper study resource allocation mech-anism design from the point of view of computational efficiency, and design allocation schemes and algorithms which could consider computational tractability, economic ef-ficiency and social equality.In the paper, we firstly study the problem of allocating a set of indivisible goods to multiple agents without side payments. We propose a parallel elicitation-free pro-tocol for allocating indivisible goods. In every round of the allocation process, some agents will be selected(according to some parallel policy) to report their preferred ob-jects among those that remain, and every reported object will be allocated randomly to an agent reporting it. Theoretical and empirical results show that parallel protocol has advantage over previous sequential elicitation-free protocol for both utilitarian and egalitarian social welfare. We also address strategical issues. Meanwhile, we study the computational issues of dynamic mechanisms for selling multiple indivisible item-s under price rigidities. We propose a polynomial algorithm that can be used to find over-demanded sets of items, and then introduce a dynamic mechanism with rationing to discover constrained Walrasian equilibria under price rigidities in polynomial time. We also address the computation of buyers’expected profits and items’expected prices, and discuss strategical issues in the sense of expected profits.