Articles | Volume 4
https://doi.org/10.5194/adgeo-4-3-2005
https://doi.org/10.5194/adgeo-4-3-2005
09 Aug 2005
09 Aug 2005

Agents, Bayes, and Climatic Risks - a modular modelling approach

A. Haas and C. Jaeger

Abstract. When insurance firms, energy companies, governments, NGOs, and other agents strive to manage climatic risks, it is by no way clear what the aggregate outcome should and will be. As a framework for investigating this subject, we present the LAGOM model family. It is based on modules depicting learning social agents. For managing climate risks, our agents use second order probabilities and update them by means of a Bayesian mechanism while differing in priors and risk aversion. The interactions between these modules and the aggregate outcomes of their actions are implemented using further modules. The software system is implemented as a series of parallel processes using the CIAMn approach. It is possible to couple modules irrespective of the language they are written in, the operating system under which they are run, and the physical location of the machine.