Bottom-up approaches to building agent-based models: discussing the need for a platform

This paper aims to help and guide the novice or experienced programmer during the development of agent-based models. In particular, a bottom-up approach to building agentbased models is reviewed. The approach leverages current state-of-the-art techniques in software development and demonstrates a design methodology, known as the ‘agile’ approach. A case study of an agent-based simulation implementing a memetic social network model of water use within an urban environment is discussed, in line with the framework drawing on concrete examples and source code. The paper concludes with a discussion of the limitations of bottom-up approaches, with suggestions and recommendations for how and why frameworks should be used.