We developed an artificial neural network as an air quality model and estimated the scope of the climate change impact on future (until 2064) summertime trends of hourly ozone concentrations at an urban air quality station in Tehran, Iran. Our developed scenarios assume that present-time emissions conditions of ozone precursors will remain constant in the future. Therefore, only the climate change impact on future ozone concentrations is investigated in this study. General Circulation Model (GCM) projections indicate more favorable climate conditions for ozone formation over the study Area in the future: the surface temperature increases over all months of the year, solar radiation increases, and precipitation decreases in future summers, and summertime daily maximum temperature increases about 1.2C to 3C until 2064. In the scenario based on present-time ozone conditions in the 2012 summer without any exceedances, the summertime exceedance days of the 8-hr ozone standard are Projected to increase in the future by about 4.2 days in the short term and about 12.3 days in the mid-term. Similarly, in the scenario based on present-time ozone conditions in the 2010 summer with 58 days of exceedance from the 8-hr ozone standard, exceedances are Projected to increase by about 4.5 days in the short term and about 14.1 days in the mid-term. Moreover, the number of Unhealthy and Very Unhealthy days in the 8-hr Air Quality Index (AQI) is also Projected to increase based on pollution scenarios of both summers.

Nonnegative matrix factorization (NMF) is a common method in data mining that have been used in different applications as a dimension reduction, classification or clustering method. Methods in alternating least square (ALS) Approach usually used to solve this non-convex minimization problem. At each step of ALS algorithms two convex least square problems should be solved, which causes high computational cost. In this paper, based on the properties of norms and orthogonal transformations we propose a framework to project NMF’ s convex sub-problems to smaller problems. This projection reduces the time of finding NMF factors. Also every method on ALS class can be used with our proposed framework.

We establish a relationship between general constrained pseudoconvex optimization problems and globally Projected dynamical systems. A corresponding novel neural network model, which is globally convergent and stable in the sense of Lyapunov, is proposed. Both theoretical and numerical Approaches are considered. Numerical simulations for three constrained nonlinear optimization problems are given to show that the numerical behaviors are in good agreement with the theoretical results.

In this paper, we study Projected non-stationary Simultaneous It-erative Reconstruction Techniques (P-SIRT). Based on algorithmic op-erators, convergence result are adjusted with Opial’ s Theorem. The advantages of P-SIRT are demonstrated on examples taken from to-mographic imaging.