Agent Based Simulation of Ecological SystemsGoogleHorde


ABSES: Agent Based Simulation of Ecological Systems

Artificial Intelligence x Ecology x Autonomous Agents x Simulation x Machine Learning

There has been a significant increase in the usage of mathematical models across almost all fields of science over the last decades, certainly related to the rapid progresses in hardware and software development. Models are now intensively used in Environmental Sciences (ES), namely in the field of Theoretic and Applied Ecology. Recently, a new field of research has been emerging from the linkage of Machine Learning (ML) techniques with ES, as reflected in several international workshops. ML may be useful for the analysis of environmental data and to help building mathematical models, among other things.

After a model is implemented, the first round of simulations is generally designed to test its internal logic against common sense knowledge. Once this task is accomplished, it is necessary to calibrate the model, i.e., to perform a second round of simulations and to tune model parameters in order to reproduce observed data. In every mathematical model, parameters regulate the behaviour of equations describing temporal and spatial changes of model state variables and their interactions.

Generally, there is some uncertainty associated with each parameter. This “tuning” procedure is a hard and “tedious” work requiring a good understanding of the effects of different parameters over the available variables. A third round of simulations is generally performed to validate the model, i.e., to compare predicted with observed results that were not used in the calibration process. Once the model is validated it may be used as a predictive tool. At this point, simulations may be set up depending on the purposes for which the model was developed. When the goal is to find some optimal solutions, a large number of trial and error simulations may be required. In the course of this processes, the modeller learns about the model and gets insight into its realism and internal dynamics.

Some of the learning processes associated with different phases of model implementation may be automated to save man power and increase the performance of some tasks, such as model calibration and the search for optimum solutions. Automatic calibration procedures already exist, based on systematic and exhaustive generation of parameter vectors and using several convergence methods. However, they require a large number of model runs and are, therefore, not applicable to very complex ecosystem models demanding large computational times. A possible alternative may be to develop a self-learning tool that simulates the learning process of the modeller about the simulated system. Regarding the usage of models to find optimal solutions to environmental problems, a similar approach may be used. In both cases, Intelligent Agents may be a sound way to implement the self-learning tools.

This project aims at developing a complete multi-agent simulation system, including an ecological simulator and an automatic Calibration Agent based on ML techniques capable of calibrating complex ecological models. Different Agents with learning and negotiation capabilities, representing the often-neglected intelligent entities present in this type of environment will also be developed. These agents will be capable of interacting with the simulation software in an intelligent manner in order to simulate the behaviour of humans and increase the simulation realism. The simulation system will be applied to ecological models of coastal ecosystems and used for aquaculture optimisation and environmental resource management.

The project main goal is to build a Multi-Agent simulation system capable of realistic simulation of costal ecological systems. Specific objectives include:

bullet Creation of calibrated models for different coastal ecosystems enabling its realistic simulation;
bullet Implementation of a user-friendly aquatic ecosystems simulation software;
bullet Project and implementation of a multi-agent simulation system enabling an ecological simulator to communicate with several modules and agents;
bullet Construction of an automatic model calibration agent based on machine learning techniques;
bullet Implementation of an on-line visualizer for ecological simulations;
bullet Construction of agents with learning and negotiation capabilities for ecological simulations;
bullet Real application to coastal ecosystem.

The project objective is also to contribute for defining fundamental ways by which correct management of environmental resources can make progress in the direction of sustainability.


Interested in this Project? Do you want to do a PhD or MSc thesis in Portugal on a related subject? Please contact me?

Pages Created and Maintained by Luís Reis [Luis Paulo Reis].
Last Update by Luís Paulo Reis: 22-06-2006.