Urban Consumption Models: Towards a Demand Driven Urban Goods and Services Model
A Sivakumar, K Arunotayanun, J W Polak, Imperial College London, UK
This paper presents an urban consumption model that drives the spatial demand for goods and services in an urban area.
The Urban Energy Systems (UES) project at Imperial College London is developing an integrated urban systems model called SynCity, which is designed to be a decision support tool that can test the effects of various alternative scenarios on the urban infrastructure and deliver optimal (or near-optimal) solutions. SynCity uniquely models both the demand and supply vectors in an urban area, the demand model being an agent-based microsimulation tool and the supply vectors being modelled variously using optimization and network design tools.
AMMUA (or the Agent-based Microsimulation Model of Urban Activities) forms the demand component of SynCity and predicts the demand for various resources such as transportation fuel, electricity, gas and heating. AMMUA itself is a system of medium-term land use and agent (household, business) evolution models, and short-term travel and resource demand models. The short term travel demand models consist of both a passenger demand component and a freight demand component. In keeping with the systematic and integrated approach of the UES project, the two demand components are being developed to be closely integrated. In other words, SynCity will model the urban flow of goods and services as it is driven by the consumption and purchasing activities of the population in the study area. To the best of our knowledge, this is the first attempt at explicitly linking consumption and travel demand. Most activity based models link activities to passenger travel demand but fail to make the connection between the activities, consumption behaviour, and the resulting travel demand due to the flow of goods and services.
In this paper, we present an urban consumption model that drives the spatial demand for goods and services in an urban area. The consumption model will predict the basket of commodities (goods and services) that a household purchases and the expenditure on each commodity as a proxy for the quantity purchased. This is developed as a multiple discrete-continuous model. This will be combined with simple probability models to predict the spatial element of the demand, whether the commodities are purchased online to be delivered at a residential location or they are purchased in store at a retail location. Activity patterns from the activity-based travel demand component of AMMUA will be used as inputs into this process. The analysis uses several different sources of data, including the Living Costs and Food Survey from the Office of National Statistics.
The urban consumption model will form the interface between the synthetic population of the study area (that is an input to the microsimulation model in SynCity) and the urban goods and services model that predicts the travel demand generated by the need for goods and services in urban areas.
Association for European Transport