Higher Moment Information Research Articles

Worst-Case Higher Moment Coherent Risk Based on Optimal Transport with Application to Distributionally Robust Portfolio Optimization

The tail risk management is of great significance in the investment process. As an extension of the asymmetric tail risk measure—Conditional Value at Risk (CVaR), higher moment coherent risk (HMCR) is compatible with the higher moment information (skewness and kurtosis) of probability distribution of the asset returns as well as capturing distributional asymmetry. In order to overcome the difficulties arising from the asymmetry and ambiguity of the underlying distribution, we propose the Wasserstein distributionally robust mean-HMCR portfolio optimization model based on the kernel smoothing method and optimal transport, where the ambiguity set is defined as a Wasserstein “ball” around the empirical distribution in the weighted kernel density estimation (KDE) distribution function family. Leveraging Fenchel’s duality theory, we obtain the computationally tractable DCP (difference-of-convex programming) reformulations and show that the ambiguity version preserves the asymmetry of the HMCR measure. Primary empirical test results for portfolio selection demonstrate the efficiency of the proposed model.

Higher order stationary subspace analysis

Non-stationarity in data is an ubiquitous problem in signal processing. The recent stationary subspace analysis procedure (SSA) has enabled to decompose such data into a stationary subspace and a non-stationary part respectively. Algorithmically only weak non- stationarities could be tackled by SSA. The present paper takes the conceptual step generalizing from the use of first and second moments as in SSA to higher order moments, thus defining the proposed higher order stationary subspace analysis procedure (HOSSA). The paper derives the novel procedure and shows simulations. An obvious trade-off between the necessity of estimating higher moments and the accuracy and robustness with which they can be estimated is observed. In an ideal setting of plenty of data where higher moment information is dominating our novel approach can win against standard SSA. However, with limited data, even though higher moments actually dominate the underlying data, still SSA may arrive on par.

Evaluation of an unsteady flamelet progress variable model for autoignition and flame development in compositionally stratified mixtures

The unsteady flamelet progress variable (UFPV) model has been proposed by Pitsch and Ihme [“An unsteady/flamelet progress variable method for LES of nonpremixed turbulent combustion,” AIAA Paper No. 2005-557, 2005] for modeling the averaged/filtered chemistry source terms in Reynolds averaged simulations and large eddy simulations of reacting non-premixed combustion. In the UFPV model, a look-up table of source terms is generated as a function of mixture fraction Z, scalar dissipation rate χ, and progress variable C by solving the unsteady flamelet equations. The assumption is that the unsteady flamelet represents the evolution of the reacting mixing layer in the non-premixed flame. We assess the accuracy of the model in predicting autoignition and flame development in compositionally stratified n-heptane/air mixtures using direct numerical simulations (DNS). The focus in this work is primarily on the assessment of accuracy of the probability density functions (PDFs) employed for obtaining averaged source terms. The performance of commonly employed presumed functions, such as the dirac-delta distribution function, the β distribution function, and statistically most likely distribution (SMLD) approach in approximating the shapes of the PDFs of the reactive and the conserved scalars is evaluated. For unimodal distributions, it is observed that functions that need two-moment information, e.g., the β distribution function and the SMLD approach with two-moment closure, are able to reasonably approximate the actual PDF. As the distribution becomes multimodal, higher moment information is required. Differences are observed between the ignition trends obtained from DNS and those predicted by the look-up table, especially for smaller gradients where the flamelet assumption becomes less applicable. The formulation assumes that the shape of the χ(Z) profile can be modeled by an error function which remains unchanged in the presence of heat release. We show that this assumption is not accurate.