Agent-Based Models (ABMs) have become popular in studying complex systems and supporting decision-making in policy modelling. This paper introduces an advancement in the field of ABMs by implementing a feature for real-time parameter changes within the Mesa modelling platform’s web visualisation module. We aimed to create a feature that would allow users to interact directly with a model and its parameters and observe immediate changes to the results while simulations are running. We detail the technical modifications made to Mesa’s visualisation architecture that enable this functionality, which is facilitated by establishing bidirectional communication between the model and the visualisation server. The resulting enhanced visualisation architecture within Mesa now allows users to modify parameters on the fly, providing immediate feedback and enabling a more nuanced understanding of complex dynamics as they unfold. This capability is demonstrated through several prototypical models, including the Schelling Segregation model and the El Farol Bar problem. These prototype models highlight the practical implications and advantages of this development for real-time policy simulation and decision support where rapid responses are crucial. This paper asserts that the novel developments that allow the integration of real-time parameter changes into ABMs can be utilised to improve their effectiveness and applicability in various disciplinary contexts.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Real Time Parameter Changes in Agent-Based Models with Mesa

  • Rajith Vidanaarachchi,
  • Branislava Godic,
  • Leon Booth,
  • Victoria Farrar,
  • Simone Pettigrew,
  • Jason Thompson

摘要

Agent-Based Models (ABMs) have become popular in studying complex systems and supporting decision-making in policy modelling. This paper introduces an advancement in the field of ABMs by implementing a feature for real-time parameter changes within the Mesa modelling platform’s web visualisation module. We aimed to create a feature that would allow users to interact directly with a model and its parameters and observe immediate changes to the results while simulations are running. We detail the technical modifications made to Mesa’s visualisation architecture that enable this functionality, which is facilitated by establishing bidirectional communication between the model and the visualisation server. The resulting enhanced visualisation architecture within Mesa now allows users to modify parameters on the fly, providing immediate feedback and enabling a more nuanced understanding of complex dynamics as they unfold. This capability is demonstrated through several prototypical models, including the Schelling Segregation model and the El Farol Bar problem. These prototype models highlight the practical implications and advantages of this development for real-time policy simulation and decision support where rapid responses are crucial. This paper asserts that the novel developments that allow the integration of real-time parameter changes into ABMs can be utilised to improve their effectiveness and applicability in various disciplinary contexts.