We point out a formal analogy between the Dirac equation in Majorana form and the discrete-velocity version of the Boltzmann kinetic equation. By a systematic analysis based on the theory of operator splitting, this analogy is shown to turn into a concrete and efficient computational method, providing a unified treatment of relativistic and nonrelativistic quantum mechanics. This might have potentially far-reaching implications for both classical and quantum computing, because it shows that, by splitting time along the three spatial directions, quantum information (Dirac-Majorana wave function) propagates in space-time as a classical statistical process (Boltzmann distribution).

The assessment of the ecosystem state, in terms of the functionality of the soil and of the soil/vegetation interactions is a relevant step of the procedures for monitoring the impact of human activities and climate changes on landscapes. The focus is here on the connectivity of the landscape with respect to the flux of different elements like water, sediment and fauna. The deliverable is organized into two parts: a) the first one introduces a new landscape, plants, landslides and erosion model, which exploits the Landscape Function Analysis framework to deal with the hydrogeological connectivity. As a result of the proposed methodology we provide an example of scenario analysis, applied to a BIO_SOS study site; b) the second one deals with the connectivity related to fauna fluxes. We introduce a new mathematical model, built on the metapopulation dynamics approach, to study predator-prey populations dynamics. The study has been motivated by the conservation issues related to the wolf-wild boar pair populating the Alta Murgia BIO_SOS study site.

It is generally accepted that visual perception results from the activation of a feed-forward hierarchy of areas, leading to increasingly complex representations. Here we present evidence for a fundamental role of backward projections to the occipito-temporal region for understanding conceptual object properties. The evidence is based on two studies. In the first study, using high-density EEG, we showed that during the observation of how objects are used there is an early activation of occipital and temporal areas, subsequently reaching the pole of the temporal lobe, and a late reactivation of the visual areas. In the second study, using transcranial magnetic stimulation over the occipital lobe, we showed a clear impairment in the accuracy of recognition of how objects are used during both early activation and, most importantly, late occipital reactivation. These findings represent strong neurophysiological evidence that a top-down mechanism is fundamental for understanding conceptual object properties, and suggest that a similar mechanism might be also present for other higher-order cognitive functions.

The convergence quality of the cross-entropy (CE) optimizer relies critically on the mechanism meant for randomly generating data samples, in agreement with the inference drawn in the earlier works--the fast simulated annealing (FSA) and fast evolutionary programming (FEP). Since tracing a near-global-optimum embedded on a nonconvex search space can be viewed as a rare event problem, a CE algorithm constructed using a longtailed distribution is intuitively attractive for effectively exploring the optimization landscape. Based on this supposition, a set of CE algorithms employing the Cauchy, logistic and Laplace distributions are experimentally validated in a wide range of optimization functions, which are shifted, rotated, expanded and/or composed, characterized by convex, unimodal, discontinuous, noisy and multimodal fitness landscapes. The Laplace distribution has been demonstrated to be more suitable for the CE optimization, since the samples drawn have jump-lengths long enough to elude local optima and short enough to preserve sufficient candidates in the global optimum neighborhood. Besides, a theoretical analysis has been carried out to understand the following: (i) benefits offered by the long-tailed distributions towards evasion of local optima; (ii) link between the variation in scale parameter estimate and the probability of producing candidate solutions arbitrarily close to the global optimum.

We consider a simplified 1-dimensional PDE-model describing the effect of contact inhibition in growth processes of normal and abnormal cells. Varying the value of a significant parameter, numerical tests suggest two different types of contact inhibition between the cell populations: the two populations move with constant velocity and exhibit spatial segregation, or they stop to move and regions of coexistence are formed. In order to understand the different mechanisms, we prove that there exists a segregated traveling wave solution for a unique wave speed, and we present numerical results on the ``stability" of the segregated waves. We conjecture the existence of a non-segregated standing wave for certain parameter values.

An algorithm for coupling SPH with an externalsolution is presented. The external solution can be either anotherSPH solution (possibly with different discretization) or a differentnumerical solver or an analytical solution.The interaction between the SPH solver and the externalsolution is achieved through an interface region. The interfaceregion is defined as a fixed portion of the computational domainthat provides a boundary condition for the SPH solver. A ghostfluid, composed by fully lagrangian particles (i.e. ghost particles)covering the interface region, is used to impose the boundarycondition. The ghost particle evolution, including its position, isintegrated in time according to the field of the external solution.The physical quantities of the ghost particles needed in theintegration scheme are obtained through an MLS interpolationon the field of the external solution. When a ghost particle crossesthe boundary of the interface region, entering in the SPH domain, it evolves according to the SPH governing equation. The spatial distribution of the ghost particles can become largely non-uniform due to the forcing by the external solution. Thus, a packing algorithm is applied on the ghost particles in the interface region, to guarantee a particle distribution suitable for SPH operators. Since the ghost particles can exit from the interface region, a seeding algorithm is needed to introduce new ghost-particles. The algorithm is tested on several benchmarks and with the external solutions given by other SPH solvers with different discretizations and by analytical solutions. The technique is deeply investigated in terms of accuracy, efficiency and possible applications. Finally a coupled simulation involving a finite volume solver is presented.

The turbulent flow behind a rotating marine propeller is analysed by integration of the Reynolds-Averaged Navier-Stokes Equations with both the Spalart & Allmaras (1994) eddy viscosity model and by a Detached Eddy Simulation approach (Spalart et al 1997) in order to assess advantages and limits of the two different turbulence models. As far as global quantities (like thrust and torque) are concerned, it is shown that the two methods perform equally well. On the contrary, local flow features (like the evolution of the wake or the onset of tip vortices instability) are capured by DES, whereas the eddy viscosity modelling proves to be overly dissipative.

The present work is aimed to assess the capability of a numerical code based on the solution of the Reynolds averaged Navier--Stokes Equations for the study of propeller functioning in off design conditions; this aspect is becoming of central interest in naval hydrodynamics research because of its crucial implications on design aspects and performance analysis of the vessel during its operational life. A marine propeller working in oblique flow conditions is numerically simulated by the unsteady Reynolds averaged Navier-Stokes equations (uRaNSe) and a dynamically overlapping grid approach. The test case considered is the CNR-INSEAN E779A propeller model. Two different loading conditions have been considered at different incidence angles (10° to 30°) in order to analyze the propeller performance during idealized off-design conditions, similar to those experienced during a tight manoeuvre. The main focus is on hydrodynamic loads (forces and moments) that act on a single blade, on the hub and on the complete propeller; peculiar characteristics of pressure distribution on the blade will be presented as well. Verification of the numerical computations have been asses sed by grid convergence analysis.

The flow past a rotating marine propeller is analyzed with the aim of establishing limits and capabilities and, hence, the field of applicability of different turbulence modeling approaches for this class of prob- lems. To this purpose the eddy viscosity model of Spalart and Allmaras (1994) [1] and the DES approach [2] have been used. It is shown that the RANSE method can give a very good prediction of global quan- tities such as thrust and torque, with a relatively small number of grid points. However, when the unsteady fluctuation of the flow or instability processes in the wake are of interest (for noise assessment, for instance), RANSE modeling proves to be too dissipative, as it smoothes out most of the finest flow fea- tures. On the contrary, DES modeling can track the vorticity field for a longer distance and successfully predicts the onset of instabilities in the wake, with excellent agreement with experiments.