Advances in Artificial Intelligence

Volume 2507 of the series Lecture Notes in Computer Science pp 97-107


How Planning Becomes Improvisation? — A Constraint Based Approach for Director Agents in Improvisational Systems

  • Márcia Cristina MoraesAffiliated withPPGC - Universidade Federal do Rio Grande do SulFACIN - Pontifícia Universidade Católica do Rio Grande do Sul
  • , Antônio Carlos da Rocha CostaAffiliated withPPGC - Universidade Federal do Rio Grande do SulESIN - Universidade Católica de Pelotas

* Final gross prices may vary according to local VAT.

Get Access


The aim of this paper is to explain how planning becomes improvisation for agents represented through animated characters that can interact with the user. Hayes-Roth and Doyle [10] proposed some changes in the view of intellectual skills traditionally studied as components of artificial intelligence. One of these changes is that planning becomes improvisation. They pointed out that like people in everyday life, animated characters rarely will have enough information, time, motivation, or control to plan and execute extended courses of behavior. Animated characters must improvise, engaging in flexible give-and-take interactions in the here-and-now. In this paper we present an approach to that change. We propose that planning can be understood as improvisation under external constraints. In order to show how this approach can be used, we present a multi-agent architecture for improvisational theater, focusing on the improvisational director’s processes.