Dynamic modelling of complex systems under deep uncertainty using an exploratory multi-method approach

© 2017 Proceedings - 22nd International Congress on Modelling and Simulation, MODSIM 2017. All rights reserved. The decision making of complex systems is challenging because of the presence of non-linearities and time delays in their structure and their behaviour. This decision making over the system lifetime is also challenged by the presence of deep uncertainty in the future behaviour of systems and in their surrounding environment. Traditional modelling approaches are inclined to consolidate all facts into a single ultimate model and to take a deterministic, optimal and predictive approach in decision making. However, they proved to be inadequate for coping with complexity and uncertainty challenges. We argue that an exploratory multi-method approach to modelling is needed for making effective and robust decisions for complex systems; the decisions which remain valid under a diverse range of future conditions. This paper illustrates the combined use of multi-method modelling and exploratory analysis in the support of complex systems decision making, with an application to asset acquisition and management and using the case of aircraft fleet as an illustrative example. First, a framework is introduced for the implementation of this multi-method exploratory approach in practice, and the model structure, developed for the case of aircraft fleet, is explained. We then discuss how the use of our new approach can improve the robustness of decisions in asset acquisition and management. An initial exploratory analysis is performed on the model under deep uncertainty conditions and with three design strategies: High Acquisition – Low Maintenance, Low Acquisition – High Maintenance, and Medium Acquisition – Medium Maintenance. The analysis of the results shows that investing on the maintenance capacity of an aircraft fleet could result in more average flying hours compared to more acquisition of new aircraft. However, this could cause two side-effects: a higher total (acquisition and maintenance) costs and a wider uncertainty in the future performance of the system (in terms of average flying hours and total costs).