Simulation model for self-adaptive applications in pervasive computing

Abstract

In autonomic computing, an application needs to be aware of its environment. While the term is not normally understood as being a physical environment, in pervasive computing many applications do actually need to monitor the physical environment in which they are deployed. Monitoring the environment often includes gathering information about the people working or living in this environment. Applications that self-adapt to changes in the monitored environment are known as context-aware. The environment is monitored using sensors, such as temperature, humidity, location sensors, etc., and use some form of logic to abduce a context. As the input of this context logic is environment sensor data, testing these applications usually requires deployment at a physical test location, often in a research laboratory. Our project aims to design a simulation model of contexts as a means to test the context logic of a context-aware application, by allowing sensor data to be produced from a description of contexts, i.e. the location and activities of people in this location, thereby allowing initial testing of a context-aware application without requiring physical deployment.