@inbook{23142,
  author = {Lam Bui and Axel Bender and Michael Barlow and Hussein Abbass},
  title = {Multiagent-Based Approach for Risk Analysis in Mission Capability Planning},
  abstract = {<p>In this chapter, we propose a multiagent-based approach for risk analysis in military capability planning. A hierarchical system is introduced that has two layers: an Option Production Layer (OPL) to find all possible options for the given planning problem, and a Risk Tolerance Layer (RTL) in which DMs’ acceptance of risk is evolved. The OPL uses metaheuristic techniques such as evolutionary algorithms to deal with multi-objectivity of a class of NP-hard resource investment problems, called the Mission Capability Planning Problem (MCPP), under the presence of risk factors. This problem has at least two inherent conflicting objectives: minimizing the cost of investment in resources as well as optimizing the makespan of plans. The framework allows for the addition of a risk-based objective to the problem in order to support risk assessment during the planning process. The RTL is run by a multi-agent system which simulates the risk attitudes of DM. The system determines different types of attitudes towards risk with each type applying to a sub-set of MCPP solutions. The goal of each agent is to maximize its risk tolerance levels with respect to a given subset of solutions determined in the OPL. Risk tolerance levels are used as surrogates for risk attitudes. The hierarchical system is flexible in terms of using a feedback mechanism when necessary. The RTL uses information from the OPL and can itself return some hyper-information to guide the OPL further. In a case study, we use a mission planning scenario to validate our proposal. The results from this study demonstrate the advantage of our proposed system. A diverse set of agents was found; hence different types of options can be grouped and offered to the decision-makers.</p>},
  year = {2010},
  journal = {Agent-Based Evolutionary Search},
  volume = {pp  77-96},
  chapter = {Adaptation, Learning, and Optimization},
  number = {5},
  publisher = {Springer Berlin Heidelberg},
  address = {Berlin},
  issn = {978-3-642-13425-8},
  isbn = {978-3-642-13424-1},
  doi = {10.1007/978-3-642-13425-8_4},
}