Constructing runtime models with bigraphs to address ubiquitous computing service composition volatility

In this thesis, we explore the appropriateness of the language abstractions provided by Bigraphs to construct a model at runtime to tackle the problem of volatility in a service composition running on a mobile device. Our contributions to knowledge are as follows: 1) We have shown that Bigraphs (Milner, 2009) are suitable for expressing models at runtime. 2) We have offered Bigraph language abstractions as an appropriate solution to some of the research problems posed by the models at runtime community (Aßmann et al., 2012). 3) We have discussed the general lessons learnt from using Bigraphs for a practical application such as a model at runtime. 4) We have discussed the general lessons learnt from our experiences of designing models at runtime. 5) We have implemented the model at runtime using the BPL Tool (ITU, 2011) and have experimentally studied the response times of our Bigraphical model. We have suggested appropriate enhancements for the tool based on our experiences. We present techniques to parameterize the reaction rules so that the matching algorithm of the BPL Tool returns a single match giving us the ability to dynamically program the model at runtime. We also show how to query the Bigraph structure.