ions taken from a vocabulary of qualitative values. Constraints on relationships among current and historical data, combined with experiential knowledge on the reliability of data sources, support judgments about the reliability of observed data. * Criteria and representation primitives for critical events and alerts. Any event whose occurrence may have a significant impact on process management control flow constitutes a "critical event." The actual conditions raised and communicated among process management tasks are called "6alerts"; for example, conditions leading to emergency shut-off, interrupting and resetting situation assessment, and rescheduling or aborting execution of actions. To perform the alerting function, process monitoring must maintain descriptions of critical events and an associated alert vocabulary. The system must link alerts to knowledge for selecting process management (sub)tasks that take care of them. Situation assessment requires representation primitives to describe situations, events, and their causalfunctional relationships in the process system. The extent to which they are explicitly embodied, or merely implicitly referred to in a vocabulary, depends on objectives. Both cases rely on process system models. The former case corresponds to having an explicit process system model sufficiently detailed for simulating developments and reasoning about simulations, while
[1]
Michael P. Georgeff,et al.
Procedural Expert Systems
,
1983,
IJCAI.
[2]
David Robson,et al.
Smalltalk-80: The Language and Its Implementation
,
1983
.
[3]
Barbara Hayes-Roth,et al.
A Blackboard Architecture for Control
,
1985,
Artif. Intell..
[4]
Giovanni Guida,et al.
Representing Procedural Knowledge in Expert Systems: An Application to Process Control
,
1985,
IJCAI.
[5]
Victor R. Lesser,et al.
Functionally Accurate, Cooperative Distributed Systems
,
1988,
IEEE Transactions on Systems, Man, and Cybernetics.
[6]
Lee D. Erman,et al.
The Hearsay-II Speech-Understanding System: Integrating Knowledge to Resolve Uncertainty
,
1980,
CSUR.
[7]
Jon Sticklen,et al.
Control Issues in Classificatory Diagnosis
,
1985,
IJCAI.