Evolution of blackboard control architectures

摘要

This paper examines the issues that arise in the control of blackboard systems for aplications with large and complicated search spaces by analyzing the evolution of blackboard control architectures. The authors feel that the issues addressed here apply more greatly to AI application domains involving complex multidimensional search, in which control knowledge is as important to successful problem solving as is domain knowledge. Evolution is viewed largely from the context of the Hearsay-II (HSII) speech understanding system. The appeal of the blackboard model is that it provides great flexibility in structuring problem solving. On the other hand, many of the features that are responsible for this flexibility make effective control difficult because they complicate the process of estimating the expected value of potential actions. Among the key themes in the evolution of blackboard control is the development of mechanisms that support more sophisticated goal-directed reasoning. In the basic control mechanism of HSII, control decisions could consider only the local and immediate effects of possible actions. Thus, the value of potential actions in meeting the system goals could be evaluated in only a limited manner. The development of appropriate abstractions of the intermediate state of problem solving can be used to evaluate the non-local effect of actions relative to the overall problem-solving goals. In addition, blackboard systems went from the implicit representation of goals in HSII to explicit representation of the goals that must be satisfied in order to meet the overall goals of the system. This allowed the implementation of various styles of goal-directed reasoning (e.g., subgoaling and planning) that were not supported in the basic HSII control mechanism. Other architectural mechanisms were concerned with efficiency issues. This article examines a number of different blackboard control architectures that have evolved from the basic model of HSII:HASP/SIAP's even-based control. CRYSALIS' hierarchial control, the DVMT's goal-directed architecture, the control blackboard architecture (BB1), model-based incremental planning for the DVMT, and the RESUN interpretation framework. A longer version of this paper is available as a technical report. It also includes analyses of the channelized, parameterized control loop version of the DVMT (Decker, Humphrey, & Lesser, 1989), ATOME's hybrid multistage control (Laasri, & Maitre, 1989) and CASSANDRA's distributed control (Craig, 1989).