Suchergebnisse

This paper considers a mixture double autoregressive model with two components, which can flexibly capture the features usually exhibited by many financial returns such as heteroscedasticity, large kurtosis and multimodal marginals. Bayesian method based on modern Markov Chain Monte Carlo (MCMC) technology is used to estimate the model parameters. The heteroscedasticity test problem for the underlying process is also addressed by means of Bayes factor. The performances of the proposed methods are evaluated via some simulations. It is shown that the MCMC algorithm is an effective tool to deal with the mixture model. Finally, the proposed model is applied to the S&P500 index data.set.

Part 3: Virtualization and Cloud Computing Technologies ; International audience ; Based on the characteristics of cloud—resources belonging to the same institution and independent resource pool, we proposed a model for the complex task-resource and task-task interactions in cloud by game theory, and proved the existence of Nash equilibrium in the game. In this game model, every task selects resources by itself, rather than the resources are allocated by cloud system. We propose two cloud resource allocation game models—CT-RAG and CS-RAG. A new cloud resource allocation strategy—Group Participation Game Strategy (GPGS) is proposed based on these two game models. We also find out and analyze the equilibrium state of the game with GPGS. The theory analysis shows that GPGS can reduce the total cost of the system in the condition that all tasks/subtasks are rational. Simulation compares Nash, GPGS, Opt and "Round-Robin". The results of evaluation show that the GPGS is better.