Many real-world problems require to optimize trajectories under constraints. Classical approaches are often based on optimal control methods but require an exact knowledge of the underlying dynamics and constraints, which could be challenging or even out of reach. In view of this, we leverage data-driven approaches to design a new end-to-end framework which is dynamics-free for optimized and realistic trajectories. Trajectories are here decomposed on function basis, trading the initial infinite dimension problem on a multivariate functional space for a parameter optimization problem. Then a maximum a posteriori approach which incorporates information from data is used to obtain a new penalized optimization problem. The penalized term narrows the search on a region centered on data and includes estimated features of the problem. We apply our data-driven approach to two settings in aeronautics and sailing routes optimization. The developed approach is implemented in the Python library PyRotor.

Image registration is the process of finding an alignment between two or more images so that their appearances match. It has been widely studied and applied to several fields, including medical imaging and biology, where it is related to morphometrics. In this paper, we present a construction of conformal diffeomorphisms which is based on constrained optimization. We consider a set of different penalty terms that aim to enforce conformality, based on discretizations of the Cauchy–Riemann equations and geometric principles, and demonstrate them experimentally on a variety of images.

When two trains travel along the same track in the same direction, it is a common safety requirement that the trains must be separated by at least two signals. This means that there will always be at least one clear section of track between the two trains. If the safe-separation condition is violated, then the driver of the following train must adopt a revised strategy that will enable the train to stop at the next signal if necessary. One simple way to ensure safe separation is to define a prescribed set of latest allowed section exit times for the leading train and a corresponding prescribed set of earliest allowed section entry times for the following train. We will find strategies that minimize the total tractive energy required for both trains to complete their respective journeys within the overall allowed journey times and subject to the additional prescribed section clearance times. We assume that the drivers use a discrete control mechanism and show that the optimal driving strategy for each train is defined by a sequence of approximate speedholding phases at a uniquely defined optimal driving speed on each section and that the sequence of optimal driving speeds is a decreasing sequence for the leading train and an increasing sequence for the following train. We illustrate our results by finding optimal strategies and associated speed profiles for both trains in some elementary but realistic examples.

One promising trend in digital system integration consists of boosting on-chip communication performance by means of silicon photonics, thus materializing the so-called Optical Networks-on-Chip. Among them, wavelength routing can be used to route a signal to destination by univocally associating a routing path to the wavelength of the optical carrier. Such wavelengths should be chosen so to minimize interferences among optical channels and to avoid routing faults. As a result, physical parameter selection of such networks requires the solution of complex constrained optimization problems. In previous work, published in the proceedings of the International Conference on Computer-Aided Design, we proposed and solved the problem of computing the maximum parallelism obtainable in the communication between any two endpoints while avoiding misrouting of optical signals. The underlying technology, only quickly mentioned in that paper, is Answer Set Programming. In this work, we detail the Answer Set Programming approach we used to solve such problem.

Another important design issue is to select the wavelengths of optical carriers such that they are spread across the available spectrum, in order to reduce the likelihood that, due to imperfections in the manufacturing process, unintended routing faults arise. We show how to address such problem in Constraint Logic Programming on Finite Domains.

Electronic communications, as well as other categories of interactions within social networks, exhibit bursts of activity localised in time. We adopt a self-exciting Hawkes process model for this behaviour. First we investigate parameter estimation of such processes and find that, in the parameter regime we encounter, the choice of triggering function is not as important as getting the correct parameters once a choice is made. Then we present a relaxed maximum likelihood method for filling in missing data in records of communications in social networks. Our optimisation algorithm adapts a recent curvilinear search method to handle inequality constraints and a non-vanishing derivative. Finally we demonstrate the method using a data set composed of email records from a social network based at the United States Military Academy. The method performs differently on this data and data from simulations, but the performance degrades only slightly as more information is removed. The ability to fill in large blocks of missing social network data has implications for security, surveillance, and privacy.

Dynamic network models describe many important scientific processes, from cell biology and epidemiology to sociology and finance. Estimating dynamic networks from noisy time series data is a difficult task since the number of components involved in the system is very large. As a result, the number of parameters to be estimated is typically larger than the number of observations. However, a characteristic of many real life networks is that they are sparse. For example, the molecular structure of genes make interactions with other components a highly-structured and, therefore, a sparse process. Until now, the literature has focused on static networks, which lack specific temporal interpretations.

We propose a flexible collection of ANOVA-like dynamic network models, where the user can select specific time dynamics, known presence or absence of links, and a particular autoregressive structure. We use undirected graphical models with block equality constraints on the parameters. This reduces the number of parameters, increases the accuracy of the estimates and makes interpretation of the results more relevant. We illustrate the flexibility of the method on both synthetic and real data.

Orbital-free density functional theory (OFDFT) is a quantum mechanical method in which the energy of a material depends only on the electron density and ionic positions. We examine some popular algorithms for optimizing the electron density distribution in OFDFT, explaining their suitability, benchmarking their performance, and suggesting some improvements. We start by describing the constrained optimization problem that encompasses electron density optimization. Next, we discuss the line search (including Wolfe conditions) and the nonlinear conjugate gradient and truncated Newton algorithms, as implemented in our open source OFDFT code. We finally focus on preconditioners derived from OFDFT energy functionals. Newly-derived preconditioners are successful for simulation cells of all sizes without regions of low electron-density and for small simulation cells with such regions.

In the setting ofa real Hilbert space ${\cal H}$, we investigate the asymptotic behavior, as time t goes to infinity, of trajectories of second-order evolutionequations

          ü(t) + γ$\dot{u}$(t) + ∇ϕ(u(t)) + A(u(t)) = 0,

where ∇ϕ is the gradient operator of a convexdifferentiable potential function ϕ : ${\cal H}\to \R$, A : ${\cal H}\to {\cal H}$ is a maximal monotone operator which is assumed to beλ-cocoercive, and γ > 0 is a damping parameter.Potential and non-potential effects are associated respectively to∇ϕ and A. Under condition λγ2 > 1, it is proved that each trajectory asymptotically weaklyconverges to a zero of ∇ϕ + A. This condition, whichonly involves the non-potential operator and the dampingparameter, is sharp and consistent with time rescaling. Passingfrom weak to strong convergence of the trajectories is obtained byintroducing an asymptotically vanishing Tikhonov-like regularizingterm. As special cases, we recover the asymptotic analysis of theheavy ball with friction dynamic attached to a convex potential, thesecond-order gradient-projection dynamic, and the second-orderdynamic governed by the Yosida approximation of a general maximalmonotone operator. The breadth and flexibility of the proposed framework is illustrated through applications in the areas of constrained optimization,dynamical approach to Nash equilibria for noncooperative games, and asymptotic stabilization in the case of a continuum of equilibria.

L'objectif de ce travail est de construire par POD ("Proper Orthogonal Decomposition") des modèles d'ordre réduit permettant de reproduire de manière fiable la dynamique spatio-temporelle d'un écoulement décollé. L'intérêt croissant pour ces modèles réduits de dynamique s'explique par leur utilisation potentielle dans la résolution de problèmes d'optimisation sous contraintes de grande taille, rencontrés en contrôle d'écoulements. La démarche générale consiste à remplacer le modèle détaillé de l'écoulement (les équations de Navier-Stokes), par un modèle approché, plus rapide à résoudre et contenant les caractéristiques essentielles de la dynamique. La POD, optimale au sens de la reconstruction énergétique, permet d'approximer la dynamique de l'écoulement dans un sous-espace généré par un petit nombre de modes. Différentes méthodes de calibration sont ici développées afin d'améliorer la prédiction du modèle de dimension réduite, et ainsi rendre cette approche utilisable dans le cadre d'une problématique de contrôled'écoulements. Essentiellement, ces méthodes consistent à ajouter aux équations du modèle des termes supplémentaires calculés comme solutions d'un problème de minimisation sous contraintes.

We propose a simple model for interaction between gene candidates in the two strands of bacterial DNA (deoxyribonucleic acid). Our model assumes that ‘final’ genes appear in one of the two strands, that they do not overlap (in bacteria there is only a small percentage of overlap), and that the final genes maximize the occupancy rate, which is defined to be the proportion of the genome occupied by coding zones. We are more concerned with describing the organization and distribution of genes in bacterial DNA than with the very hard problem of identifying genes. To this end, an algorithm for selecting the final genes according to the previously outlined maximization criterion is proposed. We study the graphical and probabilistic properties of the model resulting from applying the maximization procedure to a Markovian representation of the genic and intergenic zones within the DNA strands, develop theoretical bounds on the occupancy rate (which, in our view, is a rather intractable quantity), and use the model to compute quantities of relevance to the Escherichia coli genome and compare these to annotation data. Although this work focuses on genomic modelling, we point out that the proposed model is not restricted to applications in this setting. It also serves to model other resource allocation problems.

Let H be a real Hilbert space, $\Phi_1: H\to \xR$ a convex function of class ${\mathcal C}^1$ that we wish to minimize under the convexconstraint S.A classical approach consists in following the trajectories of the generalizedsteepest descent system (cf.   Brézis [CITE]) appliedto the non-smooth function  $\Phi_1+\delta_S$ . Following Antipin [1], it is also possible to use a continuous gradient-projection system.We propose here an alternative method as follows:given a smooth convex function  $\Phi_0: H\to \xR$ whose critical points coincidewith Sand a control parameter $\varepsilon:\xR_+\to \xR_+$ tending to zero,we consider the “Steepest Descent and Control” system \[(SDC) \qquad \dot{x}(t)+\nabla \Phi_0(x(t))+\varepsilon(t)\, \nabla \Phi_1(x(t))=0,\] where the control ε satisfies $\int_0^{+\infty} \varepsilon(t)\, {\rm d}t =+\infty$ . This last condition ensures that ε “slowly” tends to 0. When H is finite dimensional, we then prove that $d(x(t), {\rm argmin}\kern 0.12em_S \Phi_1) \to 0\quad (t\to +\infty),$ and we give sufficient conditions under which  $x(t) \to \bar{x}\in \,{\rm argmin}\kern 0.12em_S \Phi_1$ .We end the paper by numerical experiments allowing to compare the (SDC) system with the other systems already mentioned.

This paper studies optimal timing of inspections for units in cold stand-by. The problem is approached in continuous time. The policies developed adapt optimally to information provided by failures of the main unit and by ageing of both the standby and the main unit.

Results include a general form for the optimal policy, consisting of a recursive relation between successive inspection times, and of a renewal-type integral equation to determine the value of the objective function for a given policy. Further results are derived for the case in which the failure densities describing the process are logconcave. Then the stopping rule is easily characterized. It is also shown that the interval between inspections decreases with time, and that the optimal schedule is unique. A binary search algorithm for finding the solution is outlined.

Uniformization permits the replacement of a semi-Markov decision process (SMDP) by a Markov chain exhibiting the same average rewards for simple (non-randomized) policies. It is shown that various anomalies may occur, especially for stationary (randomized) policies; uniformization introduces virtual jumps with concomitant action changes not present in the original process. Since these lead to discrepancies in the average rewards for stationary processes, uniformization can be accepted as valid only for simple policies.

We generalize uniformization to yield consistent results for stationary policies also. These results are applied to constrained optimization of SMDP, in which stationary (randomized) policies appear naturally. The structure of optimal constrained SMDP policies can then be elucidated by studying the corresponding controlled Markov chains. Moreover, constrained SMDP optimal policy computations can be more easily implemented in discrete time, the generalized uniformization being employed to relate discrete- and continuous-time optimal constrained policies.

When the univariate twin design is extended by including parents of twins, it is possible to assess additive genetic effects in the presence of assortative mating and genotype-environment correlation, the effects of parental influence, as well as the extent of residual shared environmental influences. The analysis of data obtained in such an extended twin design can be carried out by means of constrained maximum likelihood confirmatory factor analysis. Specifically, the structural model underlying this design can be represented as a LISREL model with nonlinear constraints. This representation offers the possibility to consider extended multivariate twin designs involving common genetic and environmental factors. The proposed method will be illustrated with applications to simulated and real data.

Optimal causal policies maximizing the time-average reward over a semi-Markov decision process (SMDP), subject to a hard constraint on a time-average cost, are considered. Rewards and costs depend on the state and action, and contain running as well as switching components. It is supposed that the state space of the SMDP is finite, and the action space compact metric. The policy determines an action at each transition point of the SMDP.

Under an accessibility hypothesis, several notions of time average are equivalent. A Lagrange multiplier formulation involving a dynamic programming equation is utilized to relate the constrained optimization to an unconstrained optimization parametrized by the multiplier. This approach leads to a proof for the existence of a semi-simple optimal constrained policy. That is, there is at most one state for which the action is randomized between two possibilities; at all other states, an action is uniquely chosen for each state. Affine forms for the rewards, costs and transition probabilities further reduce the optimal constrained policy to ‘almost bang-bang’ form, in which the optimal policy is not randomized, and is bang-bang except perhaps at one state. Under the same assumptions, one can alternatively find an optimal constrained policy that is strictly bang-bang, but may be randomized at one state. Application is made to flow control of a birth-and-death process (e.g., an M/M/s queue); under certain monotonicity restrictions on the reward and cost structure the preceding results apply, and in addition there is a simple acceptance region.

This paper considers reinsurance retention limits in cases where the cedent has a choice between a pure quota-share treaty, a pure excess of loss treaty or a combination of the two. Our primary aim is to find the combination of retention limits which minimizes the skewness coefficient of the insurer's retained risk subject to constraints on the variance and the expected value of his retained risk. The results are given without specifying precisely how the excess of loss reinsurance premium is calculated. It is also shown that, depending to some extent on the constraint on the variance, the solution to the problem is a pure excess of loss treaty if the excess of loss premium is calculated using the expected value or standard deviation principle but that this need not be true if the variance principle is used.