William Hughes Undergraduate Dissertation 2014/15

Agent Based Modelling Of Financial Markets

Supervised by M.Stannett

Abstract

Trying to accurately forecast an economy has been proven difficult by many experts over the last 100 years. Economies are complex systems, consisting of many small actions made locally by individuals that affect much large factors. Traditionally, economists try to predict the future using models designed to examine the dynamics of aggregate quantities. However, more recently a newer method of forecasting called agent based modelling has been developed. Agent based models (ABM) try to simulate an economy through the interactions of many agents which have been assigned particular behavioural rules.

The aim of this project is to create an ABM that simulates how a financial stock market reacts when agents are represented as different types of investors. A program will be developed using Netlogo (an ABM environment) to create a model with higly changeable parameters. This model will also have a graphical user interface to make it easy for the user to operate.