PRISE - Pricing of Distributed Information Services



Pricing of Distributed Information Services

In a world of distributed information systems, which is represented by the internet, the efficient allocation of system resources is a critical coordination task. In contrast to traditional approaches of queueing theory and related scheduling algorithms PRISE is focussing also the economic interpretation of scheduling problems. Dynamic assignment of priorities to tasks seems to be the first step in this direction. The supply function is modelled as a markovian decision process, which implies the value function to be the price making factor. Connectionistic methods, like reinforcement learning and neural-networks, as well as methods of local search like genetic algorithms play an important role in the solution of the first project part. A second approach tries to employ methods of combinatorial auctions to solve the price dependant optimal allocation problem.