Integrated urban system modelling: methodology and case study using multi-agent systems

Abstract

Challenges facing urban planners and governments continue to mount as populations in urban areas increase, pressure on the world's resources reaches critical levels and degradation of ecosystems around the world becomes increasingly apparent. The movement towards sustainable development has been met with enthusiasm by decision-makers, although exactly how to achieve this target, or even measure progress towards it, is not entirely evident. This paper explores how complex urban systems can be modelled holistically using a multi-agent based framework, and their sustainability assessed using a systems approach. There are numerous subsystems and corresponding resources (natural, financial, human and man-made) within an urban development such as water, energy, transport, waste, economic and social systems. All of these subsystems and their interrelations can be modelled using multi-agent systems, along with effects of human behaviour, both spatially and temporally, in order to provide planners, developers and decision-makers with a better platform for understanding the complexities of the urban form. As well as the presentation of a general overview of how the complexities of urban systems can be best captured using integrated modelling techniques such as multi-agent systems, the AUSTIME methodology, "Assessment of Urban Sustainability Through Integrated Modelling and Exploration", will be briefly presented. This methodology is designed to show how quantifiable sustainability assessment, based on system resource thresholds, and multi-agent based modelling, can be integrated into a framework that can be used for decision making and management relating to policy, regulation, planning, design and development of urban systems. The framework is designed to form part of a cyclic process, such as an adaptive management and learning or total quality management cycle that can explicitly include stakeholder participation and ongoing evaluation. Specific examples of the implementation of this methodology are provided from a case study of Christie Walk, an eco-development in inner-city Adelaide, Australia. The case study involves the development of a prototype multi-agent based model coupled with a sustainability assessment framework that allows quantifiable sustainability comparisons for a range of indicators between the Christie Walk development and the larger Adelaide metropolitan area. Simulation results show that the eco-development performs significantly better than the majority of Adelaide metropolitan developments, specifically in carbon dioxide production, where the development's occupants rate below the 5th percentile of all Adelaide residents. The model is subsequently used to examine scenarios relating to changes in occupant behaviour, development infrastructure and location. Simulations comparing the relative impacts of car ownership and use behaviour, and infrastructure design changes (where air conditioners and heaters are required to regulate indoor temperatures), show that high in-house electricity use behaviour related to infrastructure changes to Christie Walk could have a greater effect on equivalent carbon dioxide production than increased car ownership and use (potentially due to a location change to an outer suburb of the Adelaide metropolitan area). This result is just one example that highlights the necessity for planners and governments to consider the relative importance and effects of all subsystems in urban areas on the overall system's sustainability before attempting to design and choose management options and plans. It is envisaged that the AUSTIME methodology and case study application will help provide an example of how integrated modelling and sustainability assessment can be built into adaptive management cycles and effectively used as a decision making tool to work towards the sustainable development of urban environments and their inhabitants.