Issue No. 06 - November-December (1997 vol. 9)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/69.649317
<p><b>Abstract</b>—This paper proposes a model for making the most opportune choice among a number of alternatives in situations involving trade-offs and uncertainty. Alternative actions are evaluated on the basis of strategic attributes, such as cost and effectiveness, and the choice is dynamically determined. The issue is addressed by proposing a set of rules (trade-off rules) representing strategic knowledge specifically oriented to solve trade-off problems. Trade-off rules are encoded in conditional sentences such as: "if P then Ci is preferred to Cj," where P specifies a condition that must be satisfied by the strategic attributes associated with two candidate alternatives, and Ci and Cj are general selection criteria such as "Select alternative with the minimum (or maximum) value of a specific attribute ATTR." In the first part of the paper, a deep analysis of various trade-off problems and various interdependencies among choices is presented. The second part proposes a general paradigm for strategic reasoning under trade-offs. The paradigm is evaluated using a diagnostic application in the medical domain. Finally, the proposal is discussed in the context of related work.</p>
Strategic reasoning, strategic knowledge, trade-off, expert systems, heuristic reasoning.
S. Mussi, "Strategic Reasoning Under Trade-Offs between Action Costs and Advantages," in IEEE Transactions on Knowledge & Data Engineering, vol. 9, no. , pp. 923-940, 1997.