We present a faster implementation of the polynomial time primal simplex algorithm due to Orlin [23]. His algorithm requires O(nm min{log(nC), m log n}) pivots and O(n2 m ??n{log nC, m log n}) time. The bottleneck operations in his algorithm are performing the relabeling operations on nodes, selecting entering arcs for pivots, and performing the pivots. We show how to speed up these operations so as to yield an algorithm whose running time is O(nm. log n) per scaling phase. We show how to extend the dynamic-tree data-structure in order to implement these algorithms. The extension may possibly have other applications as well.

Momentum vectors of fragment ions produced by the Coulomb explosion of CO2z+ (z = 3 - 6) and CS2z+ (z = 3 - 13) in an intense laser field (~50 fs, 1 x 1015 W/cm2) are determined by the triple coincidence imaging technique. The molecular structure from symmetric and asymmetric explosion channels is reconstructed from the measured momentum vectors using a novel simplex algorithm that can be extended to study larger molecules. Physical parameters such as bend angle and bond lengths are extracted from the data and are qualitatively described using an enhanced ionization model that predicts the laser intensity required for ionization as a function of bond length using classical, over the barrier arguments. As a way of going beyond the classical model, molecular ionization is examined using a quantum-mechanical, wave function modified ADK method. The ADK model is used to calculate the ionization rates of H2, N2, and CO2 as a function of initial vibrational level of the molecules. A strong increase in the ionization rate, with vibrational level, is found for H2, while N2 and CO2 show a lesser increase. The prospects for using ionization rates as a diagnostic for vibrational level population are assessed.

Momentum vectors of fragment ions produced by the Coulomb explosion of CO2z+ (z = 3 - 6) and CS2z+ (z = 3 - 13) in an intense laser field (~50 fs, 1 x 1015 W/cm2) are determined by the triple coincidence imaging technique. The molecular structure from symmetric and asymmetric explosion channels is reconstructed from the measured momentum vectors using a novel simplex algorithm that can be extended to study larger molecules. Physical parameters such as bend angle and bond lengths are extracted from the data and are qualitatively described using an enhanced ionization model that predicts the laser intensity required for ionization as a function of bond length using classical, over the barrier arguments. As a way of going beyond the classical model, molecular ionization is examined using a quantum-mechanical, wave function modified ADK method. The ADK model is used to calculate the ionization rates of H2, N2, and CO2 as a function of initial vibrational level of the molecules. A strong increase in the ionization rate, with vibrational level, is found for H2, while N2 and CO2 show a lesser increase. The prospects for using ionization rates as a diagnostic for vibrational level population are assessed.

No description available.

No description available.

Dans nos travaux, nous nous intéressons à la commande des convertisseurs statiques à grand nombre d'interrupteurs. Le développement des topologies multi-niveaux multi-bras a ouvert l'accès aux domaines de la forte puissance et de la haute qualité harmonique. Outre cette montée en puissance, la commande spéciale de ces dispositifs permet de conférer au convertisseur des fonctionnalités avancées de plus en plus nécessaires, comme la possibilité de filtrage actif des harmoniques, la tolérance aux pannes, la gestion du réactif, les liaisons HVDC, etc. Toutefois, un plus grand nombre d'interrupteurs au sein d'une même structure de conversion se traduit par une forte croissance du nombre de variables de commande, des degrés de liberté et par une explosion combinatoire du nombre de configurations possibles. La synthèse de lois de commande suivant les approches traditionnellement conçues pour les topologies classiques, comme les méthodes de modulation vectorielle fondées sur la représentation géométrique du convertisseur, en devient rapidement fastidieuse pour les nouvelles topologies plus complexes. De plus, les interrupteurs présents en surnombre apportent des redondances fortes qui ne sont pas nécessairement exploitées, ou du moins arbitrairement. Nous proposons une nouvelle approche de commande qui se veut moins dépendante du nombre d'interrupteurs, et qui s'affranchit des limitations induites par les méthodes de modulation géométrique. Notre approche consiste dans un premier temps à formuler de manière algébrique des problèmes de commande qui sont généralement sous-déterminés, témoignant de la présence de redondances ou degrés de liberté, et contraints, car tenant compte des limitations propres aux rapports cycliques. De manière intéressante, ces problèmes offrent une similarité avec les problèmes dits d'allocation de commande rencontrés en aéronautique, en marine ou en robotique. Dans un second temps, dans le but de fournir à chaque période de découpage une solution de commande unique et optimisée, nous concevons de nouvelles méthodes d'allocation pour les convertisseurs statiques fondées sur l'optimisation numérique en ligne à partir de techniques d'optimisation linéaire. En conséquence, les rapports cycliques sont automatiquement optimisés pour satisfaire aux références de tension tout en respectant les saturations et en exploitant les redondances disponibles selon l'état actuel du convertisseur. Nous mettons en lumière les propriétés naturellement offertes par nos méthodes. Notamment, toutes nos solutions de modulation étendent de manière maximale la zone de linéarité du convertisseur. Nous proposons des méthodes d'allocation pour la commande en tension ou en courant de topologies variées : l'onduleur quatre bras deux niveaux, l'onduleur multicellulaire à condensateurs flottants, l'onduleur modulaire multi-niveaux. Concernant les convertisseurs multicellulaires, nos méthodes d'allocation utilisent automatiquement les degrés de liberté disponible pour fournir un équilibrage actif très rapide des tensions de condensateurs flottants. Aussi, grâce à la formulation algébrique des contraintes de commande, nos algorithmes peuvent prendre en compte un défaut sur un interrupteur pour conférer au convertisseur une propriété de tolérance aux fautes du point de vue de la commande. / In our works, we are interested in control of high-switch-count power converters. The development of multileg, multilevel converters has opened the access to high power and high harmonic quality. The special control of these devices brings to the converter advanced abilities that are more and more requested nowadays, like active harmonic filtering, fault tolerance, active and reactive power transfer, High Voltage Direct Current (HVDC) links, etc. However, a higher number of switches in a conversion structure leads to a higher number of control variables, as well as more redundancies and a combinatorial explosion of the number of possible configurations. The development of control laws resulting from approaches traditionally designed for classical topologies, as for space vector modulation methods, becomes harder for new, much complex topologies. Moreover, the too many available switches bring strong control redundancies that are not necessarily exploited, at least arbitrarily. We propose a new control approach that is expected to be less dependent on the number of switches, and that does not suffer from limitations proper to geometrical modulation methods. Firstly, our approach consists in the algebraic formulation of control problems that are generally under-determined, highlighting the presence of redundancies and degrees of freedom, and constrained, because control limitations are taken into account. Interestingly, a connection can be highlighted to the so-called control allocation problem in flight control, robotics, or marine applications. Secondly, in order to compute a unique and optimized control solution at each switching period, we develop new control allocation methods for power converters based on on-line numerical optimization using linear programming techniques. Consequently, duty cycles are automatically optimized to satisfy voltage references while respecting saturations and exploiting available redundancies depending on the state of the converter. We highlight the properties naturally offered by our methods. In particular, all modulation solutions yield a maximized extension of the linearity range of the converter. We propose control allocation methods for the voltage or current control of many topologies: the four-leg two-level inverter, the multicellular flying capacitor inverter, the modular multilevel inverter.

No description available.

The aim of the thesis was to design the transformation of production using the simplex algorithm. My thesis is focus on the transformation of the original Bachelor thesis, for which I created the initial production of the holding, with a focus on plant and animal production. The transformation concentrates on the conversion of the milk's production to production focused on breeding cows without market milk production. The base stone was redoing the plant production. Plant production has been used not only for the production of custom feeding, but for the sale of plant commodities on the market to increase profits. The new production focuses on the cultivation of the crop production purely for their own consumption as feed for livestock. All profits create livestock production. The intention was therefore to transform the plant and livestock production, so as to obtain the most effective new way to use areas, for a change of the production.