Project leader: Ing. Miroslav Kárný, DrSc.

Department: AS

Supported by (ID): AVČR1ET100750401

Duration: 2004-2007

Publications at UTIA: list

Abstract:

Aim 1 Development of implementable theory of Bayesian adaptive distributed decision making (DM) with multiple participants and multiple-criteria. It will provide: Fully probabilistic design (FPD) of adaptive strategies respecting changing environment and group aims. Methodology of combining experience, observed data, statistics and individual aims. Aim 2 Transformation of the theory into a set of generic, easy-to-tailor algorithms implemented in an open software system. It will contain algorithms for: dynamic mixtures supporting FPD of distributed DM; universal approximation property and relative simplicity of their tailoring motivate this choice; automatic translation of technical knowledge into probabilistic description and solutions of FPD; and software for algorithmic development and its transfer to industry verified by creating an industrial version.