A computational experiment-based evaluation method for context-aware services in complicated environment

One context-aware service system can be implemented by various customized service strategies, by which the performance achieved can be significantly different. In particular, incorrect or non-real time context-aware service would not work well and even cause disastrous consequences in some special and complicated conditions. Therefore, performance evaluation of different service strategies in complicated environments is important. Due to the diversity of context events, and the economic, legal, and ethical limitations, however, it is rather difficult or even impossible for traditional methods to conduct comprehensive evaluation for various service strategies. In this paper, we propose a computational experiment-based evaluation method, which mainly consists of three parts: customization of context-aware services, construction of experiment system, and performance evaluation of service strategy. Furthermore, one application of this method in coal mine industry is investigated specially. Alarm and evacuation services are taken as evaluation objects, and their implementation strategies are compared by designing different ``disaster scenes'' on our experiment platform. Experiment results show the effectiveness of the proposed method.