This paper deals with variable selection in regression and binary classificationframeworks. It proposes an automatic and exhaustive procedure which relies on the use ofthe CART algorithm and on model selection via penalization. This work, of theoreticalnature, aims at determining adequate penalties, i.e. penalties whichallow achievement of oracle type inequalities justifying the performance of the proposedprocedure. Since the exhaustive procedure cannot be realized when the number of variablesis too large, a more practical procedure is also proposed and still theoreticallyvalidated. A simulation study completes the theoretical results.

Subsystems of rotary-wing aircrafts, such as helicopters for instance, are strongly interrelated due to their intrinsic specificities. Convergence to a feasible design is then not ensured and implies an iterative process. Moreover, rotorcrafts must cover a much wider range of missions than their fixed-wing counterparts. For those reasons a correctly sized rotorcraft is difficult to obtain and finding the best design for a defined set of missions needs numerous iterations. This article presents the application of a multi-objective optimization approach from the predesign stage. The standard predesign approach has been reformulated to highlight sizing constraints and three strategies are then proposed to solve those constraints: the first one is the basic transcription of the standard predesign approach; the second one leaves the problem solving to the genetic optimizer through penalization; the third one is a hybrid of both previous methods based on a constraint repairing approach. Those strategies also involved the adaptation of the helicopter modelling. Here, the focus is on two components of that new model: namely the main rotor polar and the weight assessment model.

The present paper represents a continuation of Sofonea and Matei's paper (Sofonea, M. and Matei, A. (2011) History-dependent quasivariational inequalities arising in contact mechanics. Eur. J. Appl. Math. 22, 471–491). There a new class of variational inequalities involving history-dependent operators was considered, an abstract existence and uniqueness result was proved and it was completed with a regularity result. Moreover, these results were used in the analysis of various frictional and frictionless models of contact. In this current paper we present a penalization method in the study of such inequalities. We start with an example which motivates our study; it concerns a mathematical model which describes the quasistatic contact between a viscoelastic body and a foundation; the material's behaviour is modelled with a constitutive law with long memory, the contact is frictionless and is modelled with a multivalued normal compliance condition and unilateral constraint. Then we introduce the abstract variational inequalities together with their penalizations. We prove the unique solvability of the penalized problems and the convergence of their solutions to the solution of the original problem, as the penalization parameter converges to zero. Finally, we turn back to our contact model, apply our abstract results in the study of this problem and provide their mechanical interpretation.

In this work we study a fully discrete mixed scheme, based on continuous finite elementsin space and a linear semi-implicit first-order integration in time, approximating anEricksen–Leslie nematic liquid crystal model by means of aGinzburg–Landau penalized problem. Conditional stability of this schemeis proved via a discrete version of the energy law satisfied by thecontinuous problem, and conditional convergence towards generalized Young measure-valuedsolutions to the Ericksen–Leslie problem is showed when the discreteparameters (in time and space) and the penalty parameter go to zero at the same time.Finally, we will show some numerical experiences for a phenomenon of annihilation ofsingularities.

The magnetization of a ferromagnetic sample solves anon-convex variational problem, where its relaxation by convexifyingthe energy density resolves relevantmacroscopic information. The numerical analysis of the relaxed modelhas to deal with a constrained convexbut degenerated, nonlocal energy functional in mixed formulation formagnetic potential u and magnetization m.In [C. Carstensen and A. Prohl, Numer. Math.90(2001) 65–99], the conforming P1 - (P0)d -element in d=2,3 spatialdimensions is shown to lead toan ill-posed discrete problem in relaxed micromagnetism, and suboptimalconvergence.This observation motivated anon-conforming finite element method which leads toa well-posed discrete problem, with solutions converging atoptimal rate.In this work, we provide both an a priori and a posteriori error analysis for twostabilized conforming methods which account for inter-element jumps of thepiecewise constant magnetization.Both methods converge at optimal rate;the new approach is applied to a macroscopic nonstationary ferromagnetic model [M. Kružík and A. Prohl, Adv. Math. Sci. Appl. 14 (2004) 665–681 – M. Kružík and T. Roubíček, Z. Angew. Math. Phys. 55 (2004) 159–182 ].

Abduction, first proposed in the setting of classical logics, has been studied with growing interest in the logic programming area during the last years. In this paper we study abduction with penalization in the logic programming framework. This form of abductive reasoning, which has not been previously analyzed in logic programming, turns out to represent several relevant problems, including optimization problems, very naturally. We define a formal model for abduction with penalization over logic programs, which extends the abductive framework proposed by Kakas and Mancarella. We address knowledge representation issues, encoding a number of problems in our abductive framework. In particular, we consider some relevant problems, taken from different domains, ranging from optimization theory to diagnosis and planning; their encodings turn out to be simple and elegant in our formalism. We thoroughly analyze the computational complexity of the main problems arising in the context of abduction with penalization from logic programs. Finally, we implement a system supporting the proposed abductive framework on top of the DLV engine. To this end, we design a translation from abduction problems with penalties into logic programs with weak constraints. We prove that this approach is sound and complete.