SBDO: A New Robust Approach to Dynamic Distributed Constraint Optimisation
- Cite this paper as:
- Billiau G., Chang C.F., Ghose A. (2012) SBDO: A New Robust Approach to Dynamic Distributed Constraint Optimisation. In: Desai N., Liu A., Winikoff M. (eds) Principles and Practice of Multi-Agent Systems. PRIMA 2010. Lecture Notes in Computer Science, vol 7057. Springer, Berlin, Heidelberg
Dynamic distributed constraint optimisation problems are a very effective tool for solving multi-agent problems. However they require protocols for agents to collaborate in optimising shared objectives in a decentralised manner without necessarily revealing all of their private constraints. In this paper, we present the details of the Support-Based Distributed Optimisation (SBDO) algorithm for solving dynamic distributed constraint optimisation problems. This algorithm is complete wrt hard constraints but not wrt objectives. Furthermore, we show that SBDO is completely asynchronous, sound and fault tolerant. Finally, we evaluate the performance of SDBO with respect to DynCOAA for DynDCOP and ADOPT, DPOP for DCOP. The results highlight that in general, SBDO out performs these algorithms on criteria such as time, solution quality, number of messages, non-concurrent constraint checks and memory usage.
Unable to display preview. Download preview PDF.