Wireless sensor networks (WSNs) consist of numerous sensor nodes with several possible configurations for each node. As there are a lot of nodes in a typical WSN, each with its own set of configurations, the number of configurations for the network as a whole is huge and the design space is extremely large. The configuration of a WSN has a strong effect on the quality of services of running applications and the performance of the WSN. Multi-objective evolutionary algorithms (EAs) are well suited to explore the trade-offs in a WSN design space. However, an EA has many configuration parameters in itself. This paper presents several guidelines for configuring a multi-objective EA for design space exploration, given a specification of the WSN to be configured and a time budget available for analysis. We demonstrate the effectiveness of these guidelines on a specific type of WSN that uses a gossip strategy for disseminating data over the network.
|Title of host publication||Proceedings of 4th ACM International Workshop on Performance Monitoring, Measurement and Evaluation of Heterogeneous Wireless and Wired Networks, PM2HW2N 2009, 26 October 2009, tenerife, spain|
|Place of Publication||New York|
|Publisher||Association for Computing Machinery, Inc|
|Publication status||Published - 2009|