© The Author 2009. Published by Oxford University Press on behalf of The British Computer Society. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org Advance Access publication on June 24, 2009 doi:10.1093/comjnl/bxp061 Agent Support for Policy-Driven Collaborative Mission Planning Katia Sycara 2 , Timothy J. Norman 2 , Joseph A. Giampapa 1 , Martin J. Kollingbaum 1, , Chris Burnett 2 , Daniele Masato 2 , Mairi McCallum 2 and Michael H. Strub 3,4 1 Robotics Institute, Carnegie Mellon University, 5000 Forbes Ave, Pittsburgh PA 15213, USA 2 Department of Computing Science, University of Aberdeen, Aberdeen AB24 3UE, UK 3 DSTL, Porton Down, Salisbury, Wiltshire SP4 0JQ, UK Corresponding author: mkolling@cs.cmu.edu In this paper, we describe how agents can support collaborative planning within international coalitions, formed in an ad hoc fashion as a response to military and humanitarian crises. As these coalitions are formed rapidly and without much lead time or co-training, human planners may be required to observe a plethora of policies that direct their planning effort. In a series of experiments, we show how agents can support human planners, ease their cognitive burden by giving advice on the correct use of policies and catch possible violations. The experiments show that agents can effectively prevent policy violations with no significant extra cost. Keywords: collaborative mission planning; decision support; agent aiding; policies; norms Received 4 October 2008; revised 17 April 2009 Handling editor: Erol Gelenbe 1. INTRODUCTION In order to respond rapidly and effectively to military and humanitarian crises, the formation of international coalitions by different countries and organizations is regarded as a necessity. Coalitions—an organizational form whose members engage in collaborative activities—are typically motivated by the fact that no single nation or humanitarian relief organization has all the necessary capabilities or resources to undertake particular tasks alone. Planning collaborative missions for responding effectively to crisis situations is a complicated task that requires mission planners and coordinating personnel to observe a plethora of rules or constraining policies in their attempt to deploy resources and troops in an efficient manner. Characteristically, these coalitions are formed rapidly and without much lead time or co-training. This gives rise to a set of problems. First, although coalition partners intend to collaborate for the fulfillment of common goals, they are also assumed to have individual goals whose fulfillment they are pursuing as well. In other words, coalition members are self- interested, although it is assumed that by being part of the coalition, both the individual utility and also the group utility 4 On secondment from USARL. are increased. Second, and partially as a result of the existence of individual goals, there is varying trust among the members of a coalition. Third, coalition members typically operate under different policies that range from security policies to policies about how to conduct their missions (e.g. different rules of engagement in military coalitions). The policies of different coalition partners may be in conflict and, therefore, hamper the collaborative planning of missions. In this regard, coalitions clearly differ from teams: in teamwork [1], the assumption is that team members only hold common goals and that they engage in collaborative planning and execution in pursuit of these common goals. Because of the lack of self-interest, as observable in coalitions, issues of trust and policy conflicts do not usually occur in teams. One of the most crucial challenges for coalitions is how to construct joint plans in the presence of self-interest, individual goals and diverse policies, especially in time-stressed situations, where coalition members have only limited co-training for recognizing and resolving their differences. Among others, we are mainly confronted with the following issues: in mission-critical scenarios, coalitions consisting of members from different countries or organizations are formed rapidly and without much lead time; The Computer Journal, Vol. 53 No. 5, 2010 at University of Aberdeen on June 23, 2011 comjnl.oxfordjournals.org Downloaded from