ISBN-13: 9783836452717 / Angielski / Miękka / 2007 / 96 str.
ISBN-13: 9783836452717 / Angielski / Miękka / 2007 / 96 str.
FLUX is a programming system for reasoning agents based on the fluentcalculus. FLUX agents maintain an internal model of their environment,enabling them to represent partial knowledge. This ability gives rise topowerful reasoning techniques such as planning. Conventionally, the strategysuch an agents executes in order to achieve a certain goal is static and needsto be defined in advance. Agents which are able to derive a strategy on theirown, using experience and domain knowledge, can provide a way to tacklereal world problems that change unexpectedly or for which no solution isknown in advance. In this work, we focus on learning a strategy able to usethe reasoning capabilities offered by systems such as FLUX. We present atechnique to decompose a problem into smaller subproblems using heuristicsderived from a given domain description. On this basis, subprograms arelearnt in different hypothesis spaces, which correspond to different reasoningtechniques. With the help of a grammar inducer, an inductive inferencemachine composes the found subprograms into the final strategy. Theapplication of this work is still restricted to scenarios where a simulator isavailable.