Author: "Romulus-Catalin Damaceanu"

Agent-Based Computational Economics Using NetLogo

Personal Book: US $49 Special Offer (PDF + Printed Copy): US $143
Printed Copy: US $119
Library Book: US $196
ISBN: 978-1-60805-638-5
eISBN: 978-1-60805-489-3 (Online)
DOI: 10.2174/97816080548931130101

Introduction

Agent-based Computational Economics using NetLogo explores how researchers can create, use and implement multi-agent computational models in Economics by using NetLogo software platform. Problems of economic science can be solved using multi-agent modelling (MAM). This technique uses a computer model to simulate the actions and interactions of autonomous entities in a network, in order to analyze the effects on the entire economic system. MAM combines elements of game theory, complex systems, emergence and evolutionary programming. The Monte Carlo method is also used in this e-book to introduce random elements. The 11 models presented in this text simulate the simultaneous operations of several agents in an attempt to recreate and predict complex economic phenomena.

This e-book explains the topic in a systematic manner, starting with an introduction for readers followed subsequently by methodology and implementation using NetLogo. The volume ends with conclusions based on the results of the experiments presented.

The e-book is intended as a concise and vital resource for economists, applied mathematicians, social sciences scientists, systems analysts, operations researchers and numerical analysts.

Preface

Agent-based Computational Economics applies an interdisciplinary approach that combines knowledge from Agent-based Computational Modelling and Economics with the scope to observe, analyze and discuss the evolution of an economic system composed by intelligent agents. From 1990 till now, agent-based computational modelling has gained much attention but despite this phenomenon a fundamental issue is provoked by the next key question: Can agent-based computational model represent economic reality? To answer this, we must say that any agent-based model is a model of a real economic system that takes input data and creates output data by running computer experiments.

The development of theory and applications of agent-based computational models determined in the last years has brought a real revolution regarding the modelling of complex social systems. Presently, we can say that there is a real confrontation between the adepts of equation-based modelling and of agent-based modelling. A similar "war" was encountered at the end of 19th century and the beginning of 20th century between classical economic school and neoclassical one that lasted till the middle of 20th century with the victory of neoclassical school of Economics. Starting from 1990, Agent-based Economics started an offensive movement in order to obtain important position in the system of mainstream Economics.

The confrontation between these two modelling techniques can be won by agent-based modelling only if there lies a clear strategic plan that must be followed by the scientists that are adepts of this specific technique. In our opinion, this strategic plan must contain the following objectives:

  1. Creating a general economic theory starting from the concept of agent considered as an entity that encapsulates parameters, variables, procedures and other sub-agents;
  2. Using this general theory, there must be created a set of agent-based models feasible for analyzing different aspects of economic reality;
  3. This set of models must be implemented using a general software platform like for example NetLogo;
  4. After implementing these models, then there must be created a set of computer games usable for training students in order to groom their abilities to be successful economic agents on real economic markets;
  5. We can draw the following features regarding the confrontation between adepts of equation-based and agent-based modelling in the field of Economics;
  6. The adepts of agent-base modelling are fully offensive and this reality is proved by the increasing number of papers, books, software items that deal with the subject of Agent-based Computational Economics;
  7. The adepts of equation-based modelling are in defensive positions and try to sustain their theories but the evolution of global economy encountered after 2008 is not in favour for such adepts;
  8. The economic theory that utilizes agent-based modelling is still in the pipeline and it is hoped that such theory will become mature in the middle of 21st century;
  9. If agent-based economic theory will gain the confrontation with equation-based theory then all the system of teaching economic sciences will dramatically change in the sense that students will use economic games implemented using agent-based modelling in order to obtain necessary abilities to be successful economic agents on real global economic market. At present, students memorize a set of economic theories that help them to pass exams; in the future, such students will have to pass "exams" implemented like economic games that will force our students to gain necessary abilities for becoming successful economic agents. This new training process will be exactly as the one used in our days for training pilots of airplanes; the pilot must pass the exam with flight simulator in order to become pilot of real airplanes. In other words, students must pass exams for administration of a simulated company implemented in the frame of an economic game in order to become active agent on real economic market.


Romulus-Catalin Damaceanu
Faculty of Economics
"Petre Andrei" University of Iasi
Romania
E-mail: romulus_catalin_damaceanu@yahoo.com

RELATED BOOKS

.Multi-Objective Optimization In Theory and Practice II: Metaheuristic Algorithms.
.Arduino and SCILAB based Projects.
.Arduino meets MATLAB: Interfacing, Programs and Simulink.
.Budget Optimization and Allocation: An Evolutionary Computing Based Model.