A strategy for decomposing large-scale energy-constrained sensor networks for system monitoring

作者：Ball D; Yan R; Licht T; Deshmukh A^*; Gao R X

来源：Production Planning & Control, 2008, 19(4): 435-447.

DOI：10.1080/09537280802034653

• system monitoring
• facility maintenance
• sensor networks
• energy efficiency
• Bayesian network decomposition
• uncertain reasoning

This paper presents a strategy for the optimal decomposition of large-scale sensor networks, and subsequent design of the underlying sectioned Bayesian Network (BN) for valid inferencing of system health under energy constraints. Maximising energy efficiency is critical in facility maintenance information systems that do not provide the sensors with a continuous energy source. Guidelines for sectioning the BN into a distributed, multi-agent architecture that yields a decomposed network design and satisfies any physical and inferential constraints are presented. The overall decomposition problem is formulated as a nonlinear network optimisation problem, with the objective of minimising the cumulative energy used by the sensor network while satisfying any structural and inferential constraints of the sectioned BN, and the physical requirements specified by the problem domain. Efficient methods for decomposing sensor networks into a sectioned BN are proposed for a sample sensor network system, and the computational results are presented.

• 出版日期2008