Speaker
Luís A. N. Amaral

Description System-level approaches have been a significant component of engineering for many decades. The marked aging of the populations in the developed world is posing enormous demands on the healthcare system, demanding novel approaches. Engineers are ideally placed to address several of the challenges we are facing. In my talk, I will discuss two cases where an engineering perspective and new technologies provide an opportunity for innovative approaches to two significant problems: demand for critical care and obesity.

Speaker
Josep Perelló

Description Accions, derivats, futurs, subhastes o llibre d’ordres cada cop més responen a extensos i variats bancs de dades que tempten als analistes per a ser estudiats i analitzats des d’un punt de vista estadístic. La riquesa en registres ha atret una part de la recerca provinent de les ciències naturals dins el marc de l’anomenada ciència de la complexitat. Els científics “naturals” com ara els físics o els biòlegs veuen el parquet dels mercats financers com un espai d’anàlisi amb uns models matemàtics molt aprop de l’experimentalitat amb un testeig i ajust constant entre model i observació. Podem entendre l’estudi dels mercats com a part de les anomenades ciències naturals? Aporta quelcom de diferent aquesta perspectiva? Els científics de la complexitat parlen d’estructures fractals o multifractals, clustering, cues llargues, superestadística, variables lentes i ràpides, no-linealitat, cascades, allaus…. La xerrada mostrarà diverses casuístiques conegudes des de la matemàtica financera o l’economia però abordades amb una altra mirada. Ens centrarem amb l’anàlisi de l’anomenada volatilitat estocàstica però llegida com un procés de difusió aleatòria i proposarem idees alternatives al concepte risc. O parlarem de les dades transacció a transacció com a estudi de cas per entendre millor les presa de decisió dels agents de borsa i acabarem per incidir amb l’estudi del llibre d’ordres com a escala nanoscòpica bàsica per a la formació d’un preu.

Speaker
Brian Wetton

Description Models for immiscible two-phase fluid flow in a porous medium are described. Typically, empirically fit relative permeabilities and capillary pressure functions are used to describe the macroscopic average of the behaviour in the pores. The resulting equations are of degenerate parabolic type. After a general introduction, two examples of industrial interest will be discussed. The first describes a steam-water mixture, allowing for thermal transport and phase change. Under certain conditions, a two-phase zone, in which liquid and vapour coexist, is separated from a region of only vapour by an interface. This system is a model problem for transport in hydrogen fuel cell electrodes. The second describes the removal of oil by water in soil heated by electrodes. This model describes a commercial process for environmental cleanup on small sites. In both cases, capillary pressure cannot be neglected, as in typically done in large scale reservoir and ground water simulations.

Speaker
Jordan M. Horowitz

Description The laws of thermodynamics restrict the maximum efficiency with which a thermodynamic engine can convert heat into useful work. However, this limit can be overcome with the use of feedback – that is by altering the operational protocol of the engine in response to information about its microscopic state. The reason being that the information acquired through feedback can itself be converted into work. In this talk, I will discuss recent results detailing how information enters into the thermodynamics of a system driven away from equilibrium by feedback, such as a thermodynamic engine. In particular, I will prove and analyze the second law of thermodynamics with discrete feedback, which relates information to heat, work, and irreversibility. We will see that optimal feedback processes that convert all the information into work are feedback reversible – they are indistinguishable from their time reverse. Finally building on the intuition gained from examining the second law of feedback, I will introduce a method for designing optimal thermodynamic engines with feedback, which I will illustrate with an N-particle Szilard engine.

Speaker
Theo Geisel