Multi-Objective Radiological Analysis in Real Environments

Designing systems to solve problems arising in real-world radiological scenarios is a highly challenging task due to the contextual complexities that arise. Among these are emergency response, environmental exploration, and radiological threat detection. An approach to handling problems for these applications with explicitly multi-objective formulations is advanced. This is brought into focus with investigation of a number of case studies in both natural and urban environments. These include node placement in and path planning through radioactivity-contaminated areas, radiation detection sensor network measurement update sensitivity, control schemes for multi-robot radioactive exploration in unknown environments, and adversarial analysis for an urban nuclear threat detection network. For each of these case studies, the multi-objective modeling process is undertaken along with the development of problem-specific algorithmic methods. Results take the form of outputs from multi-objective optimization algorithms; attention is given to the characteristics of both Pareto front estimations computed and samples constructed using the obtained optimal parameterizations. In the urban environment case, a novel multi-objective competitive optimization framework as well as static methods are used to produce network configurations; a quantification study is undertaken to compare performance these configurations against the adversarial behavior model.