Rozvažování v distribuovaném rozhodování

školitel: Tatiana V. Guy
e-mail: zobrazit e-mail
typ práce: dizertační práce
zaměření: MI_MM, MI_AMSM, II_SIMI
klíčová slova: deliberation, computation cost
popis: Abstrakt: Nezávisle jednající rozhodovači by měli být schopni distribuovat úsilí věnované rozvažování (scénář, v němž jeden rozhodovač přestane rozmýšlet zatím co druhý potřebuje další rozvažování může vést k nedobrému chování obou). Navíc, soutěživý scénář s několika rozhodovači od každého z nich vyžaduje sledování, jak uvažují jejich sousedé a to při omezené znalosti o nich. Title: Deliberation within distributed decision making: Why should we care? Abstract: Independently performing decision makers should be able to coordinate distribution of their deliberation efforts (scenario when a decision maker stops thinking while other needs to continue deliberation may lead to poor behaviour of both). Moreover, competitive scenario with several decision makers requires ability of each to monitor deliberation process of his neighbours under little knowledge about them. Description: The ability to think about own thinking is one of the important human’s features [4]. Long-term research in Machine Learning, Artificial Intelligence and Computer Sciences aims at solutions optimising the balance between the computational cost (time) and value of the decision-making (DM) process [2], [3]. The idea that the cost of decision making must be included into the DM process was first introduced by H.Simon [1]. The successful solution will allow an agent to stop the computation (decision-making) at the right time to optimise the overall time-dependent utility of the decision. Much of the research however has focused on a single decision maker. The PhD project is concerned with developing a formal framework for deliberation within distributed DM (e.g., cooperative robots, multi-agent systems, networks, etc.), when several agents interact, each solving own local DM problem. Independently acting agents should be able to coordinate their deliberation efforts, as each agent is uncertain about the progress made so far by its neighbours. Situation, when an agent stops thinking while other needs to continue deliberation, may lead to poor behaviour of both. Moreover, competitive scenario with several decision makers requires ability of each to monitor deliberation process of his neighbours under little knowledge about them. The aim of this PhD project is: i) to establish a theoretical understanding of deliberation (operation cost and solution time) for an agent interacting within a group; ii) make the existing normative DM theory respecting agent’s deliberation. The proposed topic of doctoral theses can be tailored to the specific PhD student, his/her skills and inclination to theoretical or algorithmic development possibly with emphasise placed on a specific application domain.
literatura: Literature: [1] Simon. H A.: A behavioral model of rational choice, In: Models of man, social and rational. Mathematical essays on rational human behavior in a social setting, John Wiley and Sons, 1957 [2] Zilberstein, S.: Metareasoning and bounded rationality. In: M. Cox and A. Raja (eds.), Metareasoning: Thinking about Thinking, MIT Press ,2011 [3] E.J. Horvitz: Rational metareasoning and compilation for optimizing decisions under bounded resources, Proceedings of Computational Intelligence, Milan, Italy, 1989 [4] G. Gigerenzer : Adaptive Thinking: Rationality in the Real World. Oxford University Press, 2000. [5] M. Kárný, T.V. Guy: Fully probabilistic control design. Systems & Control Letters, 55(4), 2006
poznámka: Good knowledge of English is essential.
naposledy změněno: 20.03.2013 11:11:21

za obsah této stránky zodpovídá: Ľubomíra Dvořáková | naposledy změněno: 12.9.2011
Trojanova 13, 120 00 Praha 2, tel. 224 358 540, pevná linka 224 923 098, fax 234 358 643
České vysoké učení technické v Praze | Fakulta jaderná a fyzikálně inženýrská | Katedra matematiky