Thesis Type: Postgraduate
Institution Of The Thesis: Uludağ Üniversitesi, Turkey
Approval Date: 2015
Thesis Language: Turkish
Student: NAGİHAN MEMİŞ
Supervisor: AZİZE GÜL EMELAbstract:
Classic single objective optimization methods are insufficient in solving problems of decision making problems which have multiple and contradictory objectives. In such a case, the use of Multiple Objective Decision Making Methods is needed. Goal Programming developed by Charnes and Cooper in 1955 is one of the well known Multiple Objective Decision Making Methods. Until today, Goal Programming have been widely used in many fields. It is not often observed in the literature the use of Goal programming in portfolio optimization.On the basis of portfolio optimization, it is aimed to perform the minimum risk for the certain level of return with the choice between alternative investments. There is one objective in this type of basic approach. However, in the portfolio selection process, decision maker may have multiple objectives apart from return and risk. The solution of this type of decision making models can not be carried out with classic single objective optimization methods. Goal Programming provide a compromise solution even in the presence of conflicting objectives by offering a flexible structure for both institutional and individual investors personel purposes. With above mentioned feature, it seems that using Goal Programming in the field of finance and especially in the solution of portfolio selection problems seems quite convenient. The purpose of this study is to perform multi objective portfolio selection also aiming companies that issue stocks performance indicator apart from stocks return and risk and analyze this portfolio's performance correspondingly. The study sample consist of stock's closing price took part BIST 100 index manufacturing industry during January 2010 - December 2013 and financial statements of companies that issue these stocks passed external audit. In generated models, profitability and market value objectives for companies is available alongside risk and return. This multiple objective optimum portolio selection models generated by Capital Asset Pricing Model is solved by using Lexicographic Goal Programming. 4 years of portfolio performances obtained by solving models is measured and Treynor Performance Measure is used in measurement. Performance solution values is compared with performance value of BIST 100 index. The comparison shows that portfolio performances provide return above the market.