GPU version of AMG preconditioner

The process of aircraft refueling has crucial impact in the performance of an airport. It is in fact of common knowledge that one of the most important indicators for benchmarking an airport is the punctuality of flights departure. To assure high results, the airplane service activities such as passengers boarding, baggage handling and aircraft refueling must not delay one another and the overall departure time.

This paper studies a scheduling problem application for the optimization of the employees used in aircrafts' refueling in a medium size airport. The problem is modelled as a particular resource leveling problem for which we provide a mixed integer mathematical formulation that we solve with CPLEX. The model allows to evaluate and analyse different scenarios that could be considered by the company in place of the current one in order to rearrange the available human resources used in refueling activity.

Background and Objective: The paper focuses on the numerical strategies to optimize a plasmid DNA delivery protocol, which combines hyaluronidase and electroporation. Methods: A well-defined continuum mechanics model of muscle porosity and advanced numerical optimization strategies have been used, to propose a substantial improvement of a pre-existing experimental protocol of DNA transfer in mice. Our work suggests that a computational model might help in the definition of innovative therapeutic procedures, thanks to the fine tuning of all the involved experimental steps.

**

-- We prove existence and uniqueness of Radon measure-valued solutions of the Cauchy problem for a first order scalar hyperbolic conservation law in one space dimension, the initial data being a finite superposition of Dirac masses and the flux being Lipschitz continuous, bounded and suciently smooth. The novelty of the paper is the introduction of a compatibility condition which, combined with standard entropy conditions, guarantees uniqueness.

A full ODE model for the transmission of cholera is investigated, includ- ing both direct and indirect transmission and a nonlinear growth for pathogens. The direct problem is preliminarily studied and characterized in terms of reproduction number, endemic and disease free equilibria. The inverse problem is then discussed in view of parameter estimation and model identification via a Least Squares Approximation approach. The procedure is applied to real data coming from the recent Yemen cholera outbreak of 2017-2018.

We introduce and analyze a new, nonlinear fourth-order regularization of forwardbackward parabolic equations. In one space dimension, under general assumptions on the potentials, which include those of Perona-Malik type, we prove existence of Radon measure-valued solutions under both natural and essential boundary conditions.