Frequently Asked Questions

These are some common questions we get, and our considered responses.

Why do I need a model?

Often you don’t need a model, but you may need a model for a range of different reasons such as:
  1. You may need to justify the business case of your actions, i.e. to quantify likely outcomes and costs.
  2. You may want to find a way to integrate your data and knowledge into decision making in a coherent manner. Models provide one way for you to achieve this.
  3. You may want to provide a way to collect organisational knowledge, data and information in the same location so that when there are staff changes, they can have easy access to this. Models can also provide a good way to get new staff up to speed about a context.
  4. You may want to manage your policies and actions in an adaptive manner. Models provide a way to explore expectations, monitor outcomes and explore why expectations and outcomes do not align. This will help you structure your investigations and data collection.

Why do you use an Agent-Based Model?

There are many types of modelling approaches that you may consider, but Agent-Based Models provide a useful approach for describing and understanding how different people are likely to respond to a decision trigger, which considers the influence of social interactions amongst a network of agents.
Agent-Based Models is an increasingly popular approach as an alternative to approaches in economics because it is better at describing diversity in a population and can embed more realistic models of decisions, such as influence from peers and social network, use of heuristics, and influence of intrinsic drivers of motivation.

Can I trust an Agent-Based Model?

Generally, the principle applies that the accuracy of your model will depend on the data and knowledge you put into it. It has been shown that statistical approaches and agent-based models have a similar capacity to predict behaviour. The main difference with an agent-based model is that it provides more flexibility when exploring hypothetical scenarios, i.e. more suitable for exploration.
We have validated our model for the water conservation context and found that we can replicate past behaviour with some level of accuracy.
There are broadly considered to be several types of validation for ABMs, with varying criticisms and benefits, but some key principles are:
  • Empiricism: Using empirical data for quantifying parameters, albeit often with acknowledgment of uncertainty that is inherent in the empirical data (i.e. you may be able to calibrate parameters within the known confidence intervals).
  • Abductive reasoning: Discounting combinations of parameter values which provide model predictions that are incompatible with empirical realisations (observed data).
  • Expert judgments: amongst the set of combined model parameters that are consistent with results of empiricism and abductive reasoning, choosing parameter values that are consistent with expert understanding of systems.
  • Sensitivity analysis: using insights from sensitivity analysis to understand how uncertainty in parameter values lead to variability in outcomes, and using this to assign uncertainty bounds on predictions.
We have applied all these principles.
References to explore further insights into validation of ABMs:
  • Brown, D. G., S. Page, et al. (2005). "Path dependence and the validation of agent-based spatial models of land use." International Journal of Geographical Information Science 19(2): 153-174.
  • Heath, B., R. Hill, et al. (2009). "A survey of agent-based modeling practices (January 1998 to July 2008)." JASSS 12(4).
  • Heckbert, S., T. Baynes, et al. (2010). Agent-based modeling in ecological economics. Annals of the New York Academy of Sciences. 1185: 39-53.
  • Klügl, F. (2008). A validation methodology for agent-based simulations. Proceedings of the ACM Symposium on Applied Computing.
  • Marks, R. E. (2007). "Validating simulation models: A general framework and four applied examples." Computational Economics 30(3): 265-290.
  • Moss, S. and B. Edmonds (2005). "Sociology and simulation: Statistical and qualitative cross-validation." American Journal of Sociology 110(4): 1095-1131.
  • Schwarz, N. and A. Ernst (2009). "Agent-based modeling of the diffusion of environmental innovations - An empirical approach." Technological Forecasting and Social Change 76(4): 497-511.
  • Windrum, P., G. Fagiolo, et al. (2007). "Empirical validation of agent-based models: Alternatives and prospects." JASSS 10(2).

What is the NetLogo modelling platform?

We use NetLogo which is a multi-agent programmable modelling environment. It is used by many tens of thousands of students, teachers and researchers worldwide. More information about the NetLogo platform can be found here,

What data do I need?

Broadly speaking, these data types are usually needed:
  • Data to understand what drives community members’ decisions. This is typically collected using a survey. We will prepare and organise the survey collection for you.
  • Data to understand what drives influencing agents’ and decision activation agents’ decisions. This can be collected either via expert knowledge (from you or directly from people in the market place) or via some sort of survey. The nature of this depends on what you will like to model so there are unfortunately no generic answers, but we are happy to discuss approach with you.
  • Past rates of adoption: to allow us to calibrate critical parameters in the model we generally will require some past rates of adoption to be available. We can operate without this, but calibration certainly helps our level of certainty in estimates. This can also be collected as part of the adaptive management of outcomes.
  • Future attributes of technology/behaviour: it may be that the performance of behaviours or technologies will change over time. If so, we will need to establish to what extent this occurs. We are happy to work with you to explore this issue.

Where can I find more information about the approach?

Here are some of the key publications from the research team:
  • Cook, S., M. Moglia, et al. (2019). Modelling Suburban Low Carbon Commuter Mode Choices. Melbourne, Low Carbon Living CRC.
  • Higgins, A., M. Syme, et al. (2014). "Forecasting uptake of retrofit packages in office building stock under government incentives." Energy Policy 65: 501-511.
  • Marquez, L., A. Higgins, et al. (2013). Understanding Influence Networks Using An Agent-Based Model of Technology Adoption for Commercial Building Retrofits. ISORAP 2013. Marrakesh, Morocco.
  • Marquez, L., J. McGregor, et al. (2017). NSW-OEH Virtual Market Simulator Demonstration Case Study. Melbourne, CSIRO.
  • McGregor, J., L. Marquez, et al. (2015). Sustainability Victoria Policy Scenario Case Study. Melbourne, CSIRO.
  • Moglia, M., S. Cook, et al. (2017). "A review of Agent-Based Modelling of technology diffusion with special reference to residential energy efficiency." Sustainable Cities and Society 31: 173-182.
  • Moglia, M., S. Cook, et al. (2018). "Promoting Water Conservation: Where to from here?" Water 10(11): 1510.
  • Moglia, M., S. Cook, et al. (2019). RP3035 Final Report: Modelling Uptake of Water Conservation and Efficiency Measures in Sydney. Melbourne, Australia, Low Carbon Living CRC.
  • Moglia, M., A. Podkalicka, et al. (2018). RP3028: Final Report: Mapping the adoption processes of energy efficient products in the residential sector. Melbourne, Australia, CSIRO.
  • Moglia, M., A. Podkalicka, et al. (2018). "An agent-based model of residential energy efficiency adoption." JASSS 21(3).
  • Tapsuwan, S., S. Cook, et al. (2018). "Willingness to pay for rainwater tank features: A post-drought analysis of Sydney water users." Water (Switzerland) 10(9).