A5065559510- Advanced Multi-Objective Optimization Algorithms
- Probabilistic and Robust Engineering Design
- Advanced Optimization Algorithms Research
- Advanced Mathematical Modeling in Engineering
- Optimal Experimental Design Methods
- Advanced Control Systems Optimization
- Advanced Thermodynamics and Statistical Mechanics
- Complex Network Analysis Techniques
- Opinion Dynamics and Social Influence
- Mathematics and Applications
- Mathematical Biology Tumor Growth
- Advanced Numerical Analysis Techniques
- Numerical methods for differential equations
- Nonlinear Photonic Systems
- Neural Networks and Applications
- Model Reduction and Neural Networks
- Metaheuristic Optimization Algorithms Research
- Quantum chaos and dynamical systems
- Reservoir Engineering and Simulation Methods
- Point processes and geometric inequalities
- Embodied and Extended Cognition
- Geometry and complex manifolds
- CO2 Sequestration and Geologic Interactions
- Geometric and Algebraic Topology
- Tree-ring climate responses
University of Salento
2024
Politecnico di Milano
2022
University of Padua
2012-2021
Centro di Ricerca in Matematica Pura ed Applicata
2005-2011
Significance For most of its path through plant bodies, water moves in conduits the wood. Plant conduction is crucial for Earth’s biogeochemical cycles, making it important to understand how natural selection shapes conduit diameters along entire lengths stems. Can mathematical modeling and global sampling explain wood ought widen from tip a trunk base? This question evolutionarily because xylem should way that keeps supply constant leaves as grows taller. Moreover, act on economy...
Lipschitz Sampling, unlike standard space lling strategies (Minimax and Maximin distance, Integrated Mean Squared Error, Eadze-Eglais, etc.) for producing good metamodels, incorporates information from output evaluation in order to estimate some sense the local complexity of function at hand. The indicator considered is a suitable denition constant. New points are proposed be evaluated where product constant by distance nearest already point maximum. Benchmarks on test functions comparison...
The temporal statistics exhibited by written correspondence appear to be media dependent, with features which have so far proven difficult characterize. We explain the origin of these difficulties disentangling role spontaneous activity from decision-based prioritizing processes in human dynamics, clocking all waiting times through each agent's `proper time' measured activity. This unveils same fundamental patterns communication across (letters, email, sms), response displaying truncated...
We propose a strategy for approximating Pareto optimal sets based on the global analysis framework proposed by Smale (Dynamical systems, New York, 1973, pp. 531-544). The method highlights and exploits underlying manifold structure of sets, optima means simplicial complexes. distinguishes hierarchy between singular set, critical set stable can handle problem superposition local fronts, occurring in general nonconvex case. Furthermore, quadratic convergence result suitable set-wise sense is...
When large volumes of fluids are removed from or injected into underground formations for, e.g., hydrocarbon and water production, [Formula: see text] storage, gas geothermal energy exploitation, monitoring surface deformations coupled to numerical modeling improves our understanding reservoir behavior. The ability accurately simulate displacements, however, is often impaired by limited information on geometry, waterdrive strength, fluid-geomechanical parameters characterizing the geologic...
In smooth and convex multiobjective optimization problems the set of Pareto optima is diffeomorphic to an $m-1$ dimensional simplex, where $m$ number objective functions. The vertices simplex are individual functions $(k-1)$-dimensional facets optimal $k$ subproblems. Such a hierarchy submanifolds geometrical object called stratification union such manifolds, in this case optima, stratified set. We discuss how these structures generalize non cases, we survey known results deduce possible...
Real-world optimization problems may involve a number of computationally expensive functions with large input variables. Metamodel-based methods can reduce the computational costs evaluating functions, but this does not dimension search domain nor mitigate curse dimensionality effects. The be reduced by functional anova decomposition involving Sobol' sensitivity indices. This approach allows one to rank decision variables according their impact on objective function values. On basis sparsity...
Abstract Deterministic global optimization algorithms like Piyavskii–Shubert, direct , ego and many more, have a recognized standing, for problems with local optima. Although single objective been extended to multiple objectives, completely deterministic nonlinear guarantees of convergence Pareto optimality are still missing. For instance, usually make use some form scalarization, which may lead incomplete representations the optimal set. Thus, all optima not be obtained, especially in...
We investigate the response function of human agents as demonstrated by written correspondence, uncovering a new universal pattern for how reactive dynamics individuals is distributed across set each agent's contacts.In long-term empirical data on email, we find that times considered separately messages to different correspondent given writer, generate family heavy-tailed distributions, which have largely same features all agents, and whose characteristic grow exponentially with rank...
In this paper, an attempt to unify two important lines of thought in applied optimization is proposed. We wish integrate the well-known (dynamic) theory Pontryagin optimal control with Pareto (of static type), involving maximization/minimization a non-trivial number functions or functionals, offers definitive theoretical device for dynamic realization objectives be optimized. The undoubtedly less known mathematical literature, even if it was studied topological and variational details (Morse...
Neural networks (NN) are a very efficient and powerful function approximation tool. Inspired by the brain structure functions, NN usually trained with backpropagation learning algorithm. A detailed benchmark on standard functions is provided, supporting in particular automatic choice of number neurons hidden layer.
In this paper we give precise asymptotic expansions and estimates of the remainder R(λ) for oscillatory integrals with non Morse phase functions, having degeneracies any order k ⩾ 2. We provide an algorithm writing down explicitly coefficients expansion analysing precisely combinatorial behaviour (Gevrey type) deriving optimal exponential decay when λ → ∞. recapture fundamental by Erdélyi (1956 Asymptotic Expansions (New York: Dover)). As it concerns estimates, seems they are novel even...
We present the exact finite reduction of a class ofnonlinearly perturbed wave equations --typically, non-linear elastic string-- based on Amann--Conley--Zehnder paradigm.By solving an inverse eigenvalue problem,we establish equivalence between spectral description derived from A--C--Z and discrete mechanical model, well definite spring--mass system. By doing so, we decrypt abstract information encoded in obtain physically sound proxy for continuous problem.
The direct algorithm has been recognized as an efficient global optimization method which few requirements of regularity and proven to be globally convergent in general cases. inspiration or used a component for many multiobjective algorithms. We propose exact genuine possible extension the multiple objectives, providing proof convergence (i.e., guarantee that infinite time becomes everywhere dense). test efficiency on nonlinear nonconvex vector function.
Abstract In this paper we propose the numerical solution of a steady‐state reaction‐diffusion problem by means application non‐local Lyapunov–Schmidt type reduction originally devised for field theory. A algorithm is developed on basis discretization differential operator simple finite differences. The eigendecomposition resulting matrix used to implement discrete version process. By new decomposed into two coupled subproblems different dimensions. large subproblem solved fixed point...