[Zurück]


Vorträge und Posterpräsentationen (ohne Tagungsband-Eintrag):

G. Feichtinger:
"Optimal control of economic processes: models of the 2nd generation";
Vortrag: IX. International Conference Approximation and Optimization in the Caribbean, San Andres Island (eingeladen); 03.03.2008.



Kurzfassung englisch:
Epidemic processes are characterised by a basic interaction mechanism. Infected individuals interact with not yet infected persons rendering them infected. The individual propensity of getting infected, i.e. a micro quantity depends on macro variables. Such diffusion mechanisms play a crucial role in various different fields like health planning, marketing, social interactions and networks, neighboring behavior, development of social norms and, last but not least, deviant behavior. The inherent nonlinearities generate multiple long-run equilibria and related complex behavior of the optimal trajectories. The thresholds separating the various basins of attraction deliver interesting insight into the 'tipping' behaviour of optimal solutions.
The purpose of this talk is to present an intertemporal cost-benefit analysis of several kinds of epidemic processes. After a short introduction into the applications of Pontryagin's maximum principle in economics and OR, we firstly present a case study, namely the US cocaine epidemic. At the strategic level, drug policy can be viewed as a resource allocation problem: how should scarce resources be divided among competing drug control programs? Drug problems are inherently dynamic, evolving over time with significant nonlinear feedback effects. Thus, one might expect the optimal mix of drug control interventions to vary over time. The lecture investigates that possibility for a couple of models of drug use with optimal control theory.
Secondly a word-of-mouth two-state compartment model is presented, the number of satisfied and not-satisfied customers being the state variables. Both groups spread messages influencing the buying behavior of potential customers in a different manner.
Thirdly, multiple equilibria and history dependence as important features of solutions in intertemporal optimization models are discussed. Multiplicity means, given the initial state of a problem, there exist multiple optimal solutions, i.e. the decision maker is indifferent which to choose. The theory is illustrated by examples from marketing, production/inventory and other fields.

Erstellt aus der Publikationsdatenbank der Technischen Universitšt Wien.