Wealth Flow Model: Online Portfolio Selection Based on Learning Wealth Flow Matrices

This article proposes a deep learning solution to the online portfolio selection problem based on learning a latent structure directly from a price time series. It introduces a novel wealth flow matrix for representing a latent structure that has special regular conditions to encode the knowledge about the relative strengths of assets in portfolios. Therefore, a wealth flow model (WFM) is proposed to learn wealth flow matrices and maximize portfolio wealth simultaneously. Compared with existing approaches, our work has several distinctive benefits: (1) the learning of wealth flow matrices makes our model more generalizable than models that only predict wealth proportion vectors, and (2) the exploitation of wealth flow matrices and the exploration of wealth growth are integrated into our deep reinforcement algorithm for the WFM. These benefits, in combination, lead to a highly-effective approach for generating reasonable investment behavior, including short-term trend following, the following of a few losers, no self-investment, and sparse portfolios. Extensive experiments on five benchmark datasets from real-world stock markets confirm the theoretical advantage of the WFM, which achieves the Pareto improvements in terms of multiple performance indicators and the steady growth of wealth over the state-of-the-art algorithms.

Groundwater Nitrate Pollution Sources Assessment for Contaminated Wellfield

Nitrates are one of the most common groundwater contaminants and they come from different sources. The paper presents a study of groundwater quality at Varaždin wellfield in the north part of Croatia. The nitrate concentration at this location has been above the maximum allowed concentration for several decades, which has made the opening of new wellfields costly. Based on the previously developed groundwater flow model, a model that covers the narrow area of the wellfield is developed. The influential zone of the observed wellfield in working conditions is determined. Based on the developed model, the sources of nitrate pollution are located, which can be generally divided into non-point and point sources. Considering the time of groundwater retention in the horizontal flow, it is concluded that the water protection zones are marked following the applicable ordinance. Based on the developed groundwater flow model at the observed wellfield, a simulation of nitrate pollution propagation by advection and dispersion processes is performed. The simulation results point out the location of the poultry landfill as the largest source of nitrate pollution. However, poultry farms, which are located in the influence area of the wellfield, also contribute significantly to the nitrate concentration at the wellfield.

A Novel Trilinear Flow Model for Capturing Dynamic Behavior of Water Flooding in Core Plug with Different Fracture Inclinations in Carbonate Reservoirs

The purpose of this study is to introduce a new three-linear flow model for capturing the dynamic behavior of water flooding with different fracture occurrences in carbonate reservoirs. Low-angle and high-angle fractures with different occurrences are usually developed in carbonate reservoirs. It is difficult to simulate the water injection development process and the law of water flooding is unclear, due to the large variation of the fracture dip. Based on the characteristics of water flooding displacement streamlines in fractured cores with different occurrences, the matrix is discretized into a number of one-dimensional linear subregions, and the channeling effect between each subregion is considered in this paper. The fractures are divided into the same number of fracture cells along with the matrix subregion, and the conduction effect between the fracture cells is considered. The fractured core injection-production system is divided into three areas of linear flow: The injected fluid flows horizontally and linearly from the matrix area at the inlet end of the core to the fracture and then linearly diverts from the fracture area. Finally, the matrix area at the outlet end of the core also presents a horizontal linear flow pattern. Thus, a trilinear flow model for water flooding oil in fractured cores with different occurrences is established. The modified BL equation is used to construct the matrix water-flooding analytical solution, and the fracture system establishes a finite-volume numerical solution, forming a high-efficiency semianalytical solution method for water-flooding BL-CVF. Compared with traditional numerical simulation methods, the accuracy is over 86%, the model is easy to construct, and the calculation efficiency is high. In addition, it can flexibly portray cracks at any dip angle, calculate various indicators of water flooding, and simulate the pressure field and saturation field, with great application effect. The research results show that the greater the fracture dip angle, the higher the oil displacement efficiency. When the fracture dip angle is above 45°, the fracture occurrence has almost no effect on the oil displacement efficiency. The water breakthrough time of through fractures is earlier than that of nonthrough fractures, and the oil displacement efficiency and injection pressure are more significantly affected by the fracture permeability. With the increase of fracture permeability, the oil displacement efficiency and the injection pressure of perforated fractured cores dropped drastically. The findings of this study can help for better understanding of the water drive law and optimizing its parameters in cores with different fracture occurrences. The three-linear flow model has strong adaptability and can accurately solve low-permeability reservoirs and high-angle fractures, but there are some errors for high-permeability reservoirs with long fractures.

Performance-based comparison of Yamada–Ota and Hamilton–Crosser hybrid nanofluid flow models with magnetic dipole impact past a stretched surface

The nanofluid flows play a vital role in many engineering processes owing to their notable industrial usage and excessive heat transfer abilities. Lately, an advanced form of nanofluids namely “hybrid nanofluids” has swapped the usual nanofluid flows to further augment the heat transfer capabilities. The objective of this envisaged model is to compare the performance of two renowned hybrid nanofluid models namely Hamilton–Crosser and Yamada–Ota. The hybrid nanoliquid (TiO2-SiC/DO) flow model is comprised of Titanium oxide (TiO2) and Silicon carbide (SiC) nanoparticles submerged into Diathermic oil (DO). The subject flow is considered over a stretched surface and is influenced by the magnetic dipole. The uniqueness of the fluid model is augmented by considering the modified Fourier law instead of the traditional Fourier law and slip conditions at the boundary. By applying the suitable similarity transformations, the system of ordinary differential equations obtained from the leading partial differential equations is handled by the MATLAB solver bvp4c package to determine the numerical solution. It is divulged that the Yamada–Ota model performs considerably better than the Hamilton–Crosser flow model as far as heat transfer capabilities are concerned. Further, the velocity reduces on increasing hydrodynamic interaction and slip parameters. It is also noted that both temperature profiles increase for higher hydrodynamic interaction and viscous dissipation parameters. The envisioned model is authenticated when compared with an already published result in a limiting case.

Network models for nonlocal traffic flow

We present a network formulation for a traffic flow model with nonlocal velocity in the flux function. The modeling framework includes suitable coupling conditions at intersections to either ensure maximum flux or distribution parameters. In particular, we focus on 1-to-1, 2-to-1 and 1-to-2 junctions. Based on an upwind type numerical scheme, we prove the maximum principle and the existence of weak solutions on networks. We also investigate the limiting behavior of the proposed models when the nonlocal influence tends to infinity. Numerical examples show the difference between the proposed coupling conditions and a comparison to the Lighthill-Whitham-Richards network model.