The Eeect of Resource Limits and Task Complexity on Collaborative Planning in Dialogue

This paper shows how agents' choice in communicative action can be designed to mitigate the eeect of their resource limits in the context of particular features of a collaborative planning task. I rst motivate a number of hypotheses about eeective language behavior based on a statistical analysis of a corpus of natural collaborative planning dialogues. These hypotheses are then tested in a dialogue testbed whose design is motivated by the corpus analysis. Experiments in the testbed examine the interaction between (1) agents' resource limits in attentional capacity and inferential capacity; (2) agents' choice in communication; and (3) features of communicative tasks that aaect task diiculty such as inferen-tial complexity, degree of belief coordination required, and tolerance for errors. The results show that good algorithms for communication must be deened relative to the agents' resource limits and the features of the task. Algorithms that are ineecient for inferentially simple, low coordination or fault-tolerant tasks are eeective when tasks require coordination or complex inferences, or are fault-intolerant. The results provide an explanation for the occurrence of utterances in human dialogues that, prima facie, appear ineecient, and provide the basis for the design of eeective algorithms for communicative choice for resource limited agents.