DR. ADOLFO LOPEZ PAREDES

”SIMAGUA: AN AGENT BASED SIMULATOR FOR WATER CYCLE MANAGEMENT IN METROPOLITAN AREAS”.

Emerging results and strategy evaluation for alternative scenarios are possible from Complex Systems simulation, with realistic and rigorous models. The practice and relevance of Production Engineering, prompted the development of commercial simulators (Witness, Arena, Siman etc) for systems and processes with well defined relations among the physical items (passive or reactive agents).

Distributed Artificial Intelligence (DAI) and in particular Multi-Agent Systems are an alternative to those simulators, with a wider scope. They allow us to model and simulate systems with higher complexity, since the agents are not limited to physical or reactive agents (objects), and path dependence and evolution of the macro-behaviour is embedded in the system. Even more, coordination and market oriented programming come up as a better alternative than central optimisation.

Water management as a public-private good is at the forefront of governments and Spain is no exception. The year 2003 was International Year for Water and the UN has declared 2005-2015 the “International Decade for Action-Water for Life” under A/RES/58/217 resolution. The EU has Water Management as a key issue in the VI Programme. And many economists and environmentalist think that in the 21 ‘st century, water will be a resource as important as oil was in the last one.

Traditional approaches to water management are short sighted, and mainly supply oriented.  The models are based on projections of historical series (time series), without systemic considerations. As their users accept, they are of very little help for policy design, although so far better than nothing.

DAI is particularly suitable to model and simulate complex systems for social choice, such as Water Cycle Management, with both demand and supply and to analyse the path dependant nature of the possible outcomes.

The project is a substantial extension of previous projects where the team has gained experience, and it will be a sound contribution to water cycle management.

We will develop an agent based simulator, flexible enough to accommodate different water management models in metropolitan areas. The simulator can accept different set-ups and be adapted to different spaces and will be linked to a GIS, to  interpret the emerging results more easily.

It will allow us to analyse different water policies jointly with alternative climate set ups or territorial changes and how these will influence a sustainable development.