The purpose of this research is to contribute to analyze market crash mechanism from the idea of a new measurement for crash risk in financial markets. Observed market crisis occurred past 20 years, the more technological development has been advanced, the more serious financial market collapsed such as the failure of Long-Term Capital Management (LTCM) taken place in summer of 1998 and Lehman collapsed in 2008. In the financial world, it has been assumed that the distributions of asset returns are all normal, however I could find it market crash should be linkage to decay or violate normality assumption. When crisis occurs, distribution becomes non-normal. Moreover recent financial technology controls return distribution that shows their investment management skills in hedge fund industry as well. The point is that if we capture timing of transition between normal state and non normal state in markets, it could be a sort of signal. For this research, correlation playing important role and calculating eigen value from correlation matrix. I observed shape of eigen value curve and used hypothetical correlation as a risk measurement parameter. Threshold shows correlation states in the market and provides an indicator.

This study presents a probability theory and a computer simulation model to analyze the risk of multiple bank failures which are caused by an asset price uctuation. The asset-side herding in the investment of banks is a potential cause of the instability of a bank system.

We attempt to visualize the collective behaviour of markets at financial crisis through cross-correlations between typical Japanese stocks by making use of multi-dimensional scaling. Then, we make a clustering of the scattered plots by minimizing the energy function of the so-called Potts spin-glasses having pair-wise interactions between spins (stocks) as correlation coefficients in stocks. We also propose a theoretical framework to predict several time-series simultaneously by using cross-correlations in financial markets. Our model system is basically described by a variant of the Ising model introduced by Kaizouji (2000). The justification and validity of our approaches are numerically examined for Japanese NIKKEI stocks around 11 March 2011, and for foreign currency exchange rates around Greek crisis in spring 2010.

The Portfolio Optimization problem is a multi multi-objective resource allocation problem where money to be allocated to the assets is the resource. The problem consists of the selection of assets from thousands of them available in the market, weigh them proper ly in the portfolio in order to minimize the risk and maximize the expected return of the investment. In our work, the main motive is to make the portfolio more realistic, apart from achieving better results. We introduce mainly: 1) The inclusion of real l ife constraints namely ? realistic transaction costs, 2) The new co co-ordinate ascent Genetic Algorithm, taking into consideration the traded volumes . We compare our results with simple GA based method and the index, and observe a noticeable improvement.

The optimization of trading rules in Foreign Exchange (FX) by using metaheuristics such as a Genetic Algorithm (GA) has been recently proposed. GA can learn trends and generate proper rules for FX. However, trends of exchange fluctuations always change by various factors. In such situations, it is difficult to apply one specific rule generated by GA for gaining benefits. This paper proposes the dynamic switching of generating rules by GA. The similarity between the learning data and the test data is calculated by correlation coefficient. This allows generated rules to adopt complex trends according to the circumstances. By conducting substantial simulations, we found that the medium correlation generated best benefits than strong correlations.