***LITE MAT***

Ett informationsblad från matematiska institutionen vid Linköpings universitet

v48 1996

Föreläsningar

Matematiska kollokviet

Fredagen den 22 november kl 13.15 talar professor Anders Melin, Lund: Om solitoner och lineariseringen av KdV-ekvationen.

Fredagen den 29 november kl 13.15 talar Jan-Olav Rönning om Generaliserade Perronträd med tillämpningar på maximalfunktioner.

Fredagen den 6 december kl 13.15 talar professor Iossif Ostrovskii, Charkov och Uppsala, om Zero distribution and growth of entire functions which are Fourier transforms of probability measures.

Fredagen den 13 december kl 13.15 talar docent Mats Andersson, Göteborg.

Lokal: MAI:s seminarierum, ing 23-25, 1 tr.

Lars Inge Hedberg

Licentiatseminarium

Fredagen den 13 december kl. 10.15 presenterar fil. mag. Hans B. Johansson, högskolan i Skövde, sin licentiatavhandling Uniform Approximation by Harmonic Functions. Opponent är docent Mats Andersson, Göteborg.

Lokal: MAI:s seminarierum, ing 23-25, 1 tr.

Lars Inge Hedberg

Seminarium Optimeringslära

Onsdagen den 27 november kl 15.15, ger Lars-Göran Mattsson, Department of Infrastructure and Planning, KTH, ett seminarium med titeln: GEV-models based on principles of random utility maximisation

Abstract:

Multinomial logit models (or more generally Generalised Extreme Value (GEV)-models) have become a standard tool in modelling discrete choice behaviour in many fields. This is particularly true for transport demand modelling of e.g. route, mode, car ownership, destination and location choices. Logit models can be derived from different assumptions. In this talk we focus on the principle of random utility maximisation. The probability that an individual chooses a particular alternative among a finite number is then identified with the probability that this alternative maximises the utility. The particular assumption that has to be made is that the utility of an alternative is additively separated into a deterministic observable term and a stochastic term, for which the distribution is supposed to be known. Depending on the choice of distribution, the class of GEV-models or other classes of discrete choice models such as the probit models will result. One advantage of interpreting GEV-models within such a micro-economic framework is that this naturally leads to an evaluation criterion, the expected utility of the chosen (utility-maximising) alternative. GEV-models possess the less well-known property that the distribution of achieved utility is invariant across alternatives. In fact this property characterises a slightly more general class of additive random utility models including GEV-models.

Lokal:: MAI:s seminarierum, ing 23 bv

Välkomna,
Jan Lundgren

Gäster

Under november och december gästas
optimeringslära av Professor David Boyce, Urban Transportation Center,
University of Illinois at Chicago, han kommer att ge följande seminarier:

Torsdagen den 28 november kl 9.15 - 12.00

Principles of Route Choice Models: Static and Dynamic Formulations

Abstact:
This seminar will explore the basic formulations of traveler choices of routes in a road network using methods of optimization and variational inequalities. Both static and dynamic cases will be examined.

The basic principle of route choice in a road network in which travel times are flow-dependent (that is congestion exists) is that for each origin-destination pair, all used routes have equal travel time and no unused route has a lower travel time. This statement can be shown to correspond to the necessary conditions for the optimum of a convex optimization problem, and directly to the elements of a variational inequality problem. This statement applies to the static case where flows are constant over a relatively long time period. It can also be seen to apply to the dynamic case where the flows are time-dependent.

In this seminar, these formulations will be studied primarily for the deterministic case in which drivers are assumed to have perfect information on the travel times prevailing on the road network. The presentation will emphasize formulations, and will not consider solution algorithms or applications in transportation planning or traffic operations.

Lokal: MAI:s seminarierum, ing 23-25, 1tr

Tisdagen den 3 december kl 13.15 - 16.00

Advanced Route Choice Models: Extensions to More Realistic Cases

Abstract:
In the sequel to the first seminar, extensions of static and dynamic models to more realistic cases will be considered. The first case is to relax the assumption that travel times are known perfectly. If random perception errors exist, or link travel times are random variables, then stochastic route choice models can be formulated for both static and dynamic cases.

Moreover, origin-destination flows may be regarded as depending on route travel times, leading to variable or elastic demand models. Finally, the relationships between link travel times and link flows may be regarded as depending on several conflicting link flows, rather than the link's own flow only. As in the first seminar, formulations with optimization and variational inequality methods will be stressed.

Lokal: MAI:s seminarierum, ing 23, bv

Välkomna,
Jan Lundgren

Presentation av examensarbete

Fredagen den 22 november kl 16.00 presenterar Oskar Enoksson (Y-tm) sitt examensarbete, inom området tillämpad matematik, Artificial diffusion and solid wall boundary conditions for a 3D multiblock Euler/Navier-Stokes solver (Multnas)

Lokal: MAI:s seminarierum

Per Weinerfelt

Mer information om MAI finns på under MAIs hemsida
Material till Lite Mat lämnas till Maud Lindström senast torsdagar kl 12.00.
Tel 013-281405, Fax 013-100746, Email: litemat@mai.liu.se
Denna sida har besökts gånger