Optimization of portfolios with and without constraints using evolutionary algorithms

• Main Author: Siddhant Gupta.
• Other Authors: Ponnuthurai Nagaratnam Suganthan
• Format: Final Year Project
• Language: English
• Published: 2009
• Subjects: DRNTU::Engineering
• Online Access: http://hdl.handle.net/10356/17904
• Institution: Nanyang Technological University

Diversification through portfolio construction has become an increasingly important tool in finance for minimizing risk associated with investment. There are two objectives that need to be optimized during portfolio constructions, thus making it a real-world multi-objective optimization problem. One quantitative approach that has generated considerable interest and extensive practice in the past few years is asset allocation. Very simply defined, asset allocation is the process of selecting a mix of asset classes and allocating capital to those assets by matching rates of return to a specified and quantifiable tolerance for risk. The idea of applying the biological principle of natural evolution to artificial systems, introduced more than three decades ago, has seen impressive growth in the past few years. Usually grouped under the term evolutionary algorithms or evolutionary computation, we find the domains of genetic algorithms, evolution strategies, evolutionary programming, and genetic programming. Evolutionary algorithms are ubiquitous nowadays, having been successfully applied to numerous problems from different domains, including optimization, automatic programming, machine learning, economics, operations research, ecology, population genetics, studies of evolution and learning, and social systems. Since the 1980s, Genetic Algorithm has been researched and used extensively to solve for these multi-objective optimization problems. Genetic Algorithm offers speed, robustness and many other interesting features that make it suitable for this kind of application. [1] In this project, Genetic Algorithm is used to construct optimal portfolio with maximum return and minimum risk. Additionally different risk models will also be studied and implemented in the portfolio optimization process. The constructed portfolio will be tested using back-testing method for its performance. Further, the project looks into the constraints of genetic algorithms. The complexity of the computation process of such algorithms and the need to develop an accurate and efficient system of portfolio optimization is solved by the use of filters which reduces the large chunk of stocks to a smaller number. Five filter processes have been developed and researched for the purpose of doing the above their results have been compared.