Satis cing Evaluation Functions The Heart of the New Design to Criteria Paradigm

Design to Criteria scheduling is the process of custom tailoring a way to achieve a high level task via actions described in a T MS model of the task to t a particular client s quality cost and duration criteria or needs At the heart of the Design to Criteria paradigm is the ability to determine how well a particular schedule or schedule abstraction ts a set of design criteria The process of measuring goodness of schedules or alternatives and determining which items are best is called evaluation Working to meet criteria is ubiquitous in the Design to Criteria scheduling system and consequently evaluation is used at every turn The new evaluation functions operate to determine a principled measurement of goodness based on relativity and proportionality Relativity is important because the objective is to make satis cing choices and the goodness of one option is relative to the other possible options Proportionality is a major concern because we do not want di erent quality cost and duration scales to skew the evaluation mechanism and because the client s criteria is described in a relative proportionalistic fashion The new evaluation functions are paired with a new criteria speci cation metaphor importance sliders The slider metaphor enables clients users or other systems to de ne the relative importance of quality cost and duration with respect to three classes of concerns raw goodness thresholds and limits and uncertainty This material is based upon work supported by the National Science Foundation under Grant No IRI the Department of the Navy O ce of the Chief of Naval Research under Grant No N and via a subcontract from Boeing Helicopter which is being supported by DARPA under the RaDEO program contract number NANB H The content of the information does not necessarily re ect the position or the policy of the Government National Science Foundation or Boeing Helicopter and no o cial endorsement should be inferred