方法对比

	Delphi Technology Forecasting	Relevance Tree Analysis
领域	Futures Foresight Studies	Futures Foresight Studies
方法族	Process / pipeline	Process / pipeline
起源年份	1964	1966
提出者	Olaf Helmer, Norman Dalkey, and colleagues at the RAND Corporation	Defense and aerospace planners (Honeywell's PATTERN program); systematized in futures research by Gordon, Helmer and colleagues
类型	Iterative anonymous expert-panel pipeline for forecasting technological events	Normative hierarchical decomposition pipeline for R&D priority-setting
开创性文献	Glenn, J. C., & Gordon, T. J. (Eds.). (2009). Futures Research Methodology, Version 3.0. The Millennium Project. ISBN: 9780981894119	Glenn, J. C., & Gordon, T. J. (Eds.). (2009). Futures Research Methodology, Version 3.0. The Millennium Project. ISBN: 9780981894119
别名	Technological Delphi, RAND Delphi Forecasting, Expert Panel Technology Forecasting, Delphi Event Forecasting	Relevance Tree Method, Relevance Number Analysis, Normative Relevance Tree, PATTERN-Style Relevance Trees
相关	3	3
摘要	Delphi technology forecasting is the original and best-known application of the Delphi method: using iterative, anonymous rounds of expert judgment with controlled statistical feedback to forecast the timing and probability of specific technological developments. Developed at the RAND Corporation by Olaf Helmer, Norman Dalkey, and colleagues, the technique was designed to harness expert opinion systematically while suppressing the social pressures of face-to-face committees — dominant personalities, bandwagon effects, and reluctance to abandon a stated position. Rather than asking a panel for a general opinion, technological Delphi asks experts to predict the year by which a well-defined development will occur, or the probability that it will occur by a given date, and then feeds back the panel's median and spread so that experts can reconsider in light of the group. Glenn and Gordon's Futures Research Methodology treats it as a foundational structured-judgment method of the field.	Relevance tree analysis is a normative forecasting method that decomposes a high-level objective into a hierarchy of sub-objectives, functions, and contributing technologies, and then assigns relevance numbers that quantify how much each branch contributes to its parent. By normalizing these numbers so that the children of every node sum to one and multiplying them down each path, the method produces an overall relevance score for every technology or task at the leaves, which ranks them by their importance to the top objective. Unlike exploratory forecasting, which projects what the future will be, relevance trees work backward from a desired goal — they are 'normative,' starting from where you want to go and identifying what must be developed to get there. Originating in defense and aerospace planning and codified in Glenn and Gordon's Futures Research Methodology, the technique remains a standard tool for research-and-development priority-setting and mission analysis.