Reference: Langlotz, C. P.; Fagan, L. M.; Tu, S. W.; Sikic, B. I.; & Shortliffe, E. H. Combining Artificial Intelligence and Decision Analysis for Automated Therapy Planning Assistance. 1986.
Abstract: Through our experience with the ONCOCIN consultation system we have identified a set of challenging planning problems to which no single existing technique readily applies. In response to these problems a computer program called ONYX has been devised which combines the artificial intelligence and decision theoretic approaches to planning. The program's planning process consists of three steps: (1) using heuristics derived from therapy planning strategies, generate a small set of plausible plans, (2) using knowledge about the structure and behavior of the human body, simulate to predict possible consequences for the patient if each plausible plan were carried out, and (3) using decision analysis, rank the plans according to how well the probabilistic simulation for each plan meets the treatment goals. The architecture explicitly manages the uncertainty inherent in many planning tasks and may have considerable utility for problems that are difficult to fit into existing frameworks.
Notes: Journal Memo.