A Combinatorial Auction Framework for Solving Decentralized Scheduling Problems
Conference Proceeding Article
Computing an optimal solution to an integer program for a realistic scheduling problem can often be very time consuming. As a result, ad-hoc and hand-crafted heuristics have become popular alternatives. These methods, however, suffer from the inability to guarantee good or optimal solutions. Our objective in this work is to leverage on prior theoretical results from the OR literature and agent technology from the AI literature to derive a computational framework that can be easily implemented for solving decentralized scheduling problems. In decentralization, there is an issue that information and control are inherently private to individual agents, even though a solution has to be jointly derived. In this paper, we are concerned about the following class of decentralized scheduling problems: 1 There is a central pool of limited resources that comprises multiple units of resources for each machine type; 1 There are multiple self-interested agents and each has to obtain resources from the central pool to solve its own scheduling problem (job shop, flow shop, etc).
Artificial Intelligence and Robotics | Business | Operations Research, Systems Engineering and Industrial Engineering
Intelligent Systems and Decision Analytics
Integration of AI and OR Techniques in Constraint Programming for Combinatorial Optimization Problems:
LAU, Hoong Chuin; Lye, Kong Wei; and NGUYEN, Viet Bang.
A Combinatorial Auction Framework for Solving Decentralized Scheduling Problems. (2008). Integration of AI and OR Techniques in Constraint Programming for Combinatorial Optimization Problems:. 5015, 333-337. Research Collection School Of Information Systems.
Available at: http://ink.library.smu.edu.sg/sis_research/364