Decentralised coordination of low-power embedded devices using the max-sum algorithm
Proceedings of the 7th international joint conference on Autonomous agents and multiagent systems - Volume 2
Effective Variants of the Max-Sum Algorithm for Radar Coordination and Scheduling
WI-IAT '11 Proceedings of the 2011 IEEE/WIC/ACM International Conferences on Web Intelligence and Intelligent Agent Technology - Volume 02
AINTEC'06 Proceedings of the Second Asian international conference on Technologies for Advanced Heterogeneous Networks
Hi-index | 0.00 |
Many message passing algorithms on graphical models include maximization operations on sums of local node function and message values from neighbors. In recent work by McAuly et al, faster maximization computation was achieved in a static environment by offline presorting of the values of local functions. However, this efficiency is only guaranteed in special cases when constraint nodes receive messages involving fewer variables than the local function. In this paper, we generalize the approach to be applicable to more general settings where offline presorting of constraint functions is not realistic and messages may involve as many variables as the constraint function. We further improve the approach in two ways, first by creating different value sets with sum values from the previous cycle and the changes in message values from the current cycle, and second by conditionally applying the technique based on a correlation measure. These new approaches with no preprocesing step obtain the expected computational complexity with an exponent of 1.5 of the possible values per node except the initial cycle which requires 2. We demonstrate the effectiveness of this approach in a distributed optimization problem involving the coordination and scheduling of radars.