Frequently Asked Questions
Q: What is a decision analyst?
A: A decision analyst is a consultant with training and experience in the logic, psychology, purposes, and methods of decision making, as well as in the analysis of decisions, whose role is to help managers invent, analyze, and select options for decisions they face, or to develop information, tools, or processes.
Q: What is decision analysis?
A: "Decision analysis is a philosophy, a body of knowledge, and a professional practice for the logical illumination of decision problems; it simultaneously considers the uncertain, dynamic, and complex consequences of a decision, as well as the assignment of value to its consequences. Many large and important problems covering the spectrum of business, government, medicine, and law have been successfully treated by decision analysis. Applications have been made to such problems as evaluating hazardous processes, research and development, and energy investment." -- Stanford University, Department of Management Science and Engineering
Q: What is a decision tree?
A: It is a central tenet of decision analysis that all relevant considerations in a decision can be assigned to one or another of four components: initial and subsequent options, possible outcomes, utility measures for outcomes, and uncertainties about random events that may affect the possible outcomes. In principle, these components can be represented fully in a decision tree. In other words, for every conceivable decision, it is theoretically possible to construct a decision tree capturing everything that a decision maker feels is relevant to the choice in question.
A decision tree is a kind of road map. Like a road map, it visually displays possible destinations (outcomes) at which you may arrive if you take one or another of the routes (initial options) immediately available to you. It also shows what you may come across on the way (events) and what later choices (subsequent options) you may have to make on the way.
A decision tree consists visually and mathematically of a network of branches corresponding to possible sequences of options and events, fanning out from one origin and reaching every consequence of possible importance. For any complex decision, it can become a very bushy tree indeed.
When certain conventions are followed in assigning probabilities and utilities, a decision tree becomes more than just a way of representing decisions. It becomes a mathematical structure for analyzing the best options to take throughout the tree and especially at the beginning of the tree, the initial option. The method of analysis is known formally as probabilistic dynamic programming, and informally as "folding back the decision tree." It is a relatively simple rote procedure.
There are several commercial software tools for analyzing decision trees. The magazine OR/MS Today regularly publishes a survey of decision analysis tools, including tools for decision trees.
The classic text on the analysis of decision trees is Decision Analysis; Introductory Lectures on Choices under Uncertainty, by Howard Raiffa of Harvard University, published in 1968. Much of the description above is quoted or adapted from the Handbook for Decision Analysis (p. 2) published by Decisions and Designs, Inc., in 1977 for the Defense Advanced Projects Agency and the Office of Naval Research, authored by Scott Barclay, Rex Brown, Clint Kelly, Cam Peterson, Larry Phillips, and Judith Selvidge.
Q: What is a decision conference?
A: Decision Conferencing is a proven process for structured planning and decision making.
• Used for strategic planning, budgeting, downsizing, system and facility design, site selection, reorganization, and more
• Ten to twenty knowledgeable, motivated stakeholders with diverse opinions, seeking a way forward
• Intensive working sessions, typically two full days
• Facilitated by IDI analysts, who guide the group to create, analyze, and reach closure on plans
• Supported with special analytical software
• Documented by IDI for the stakeholders, typically within two weeks
Conferences start with a free-wheeling exploration of the problem and the issues at stake. IDI analysts then partition the problem into its component parts and guide the group through an analysis. Participants debate the issues constructively while communicating clearly about well-defined options and values. Discussions lead to consensus on a well-informed, optimal plan for action with a broad base of understanding and support. Creates a high probability of organizational acceptance and successful implementation.
For further information see: http://www.lse.ac.uk/management/documents/WP-06-85.pdf
Q: What is a multi-attribute utility analysis?
A: Multi-criteria analysis establishes preferences between options by reference to an explicit set of objectives that the decision making body has identified, and for which it has established measurable criteria to assess the extent to which the objectives have been achieved. In simple circumstances, the process of identifying objectives and criteria may alone provide enough information for decision-makers. However, where a level of detail broadly akin to cost benefit analysis is required, multi-criteria analysis offers a number of ways of aggregating the data on individual criteria to provide indicators of the overall performance of options.
From Multi-Criteria Analysis; A Manual, p. 23, published by the Department of the Environment, Transport and the Regions, London, England, December 2000
Q: What is a Bayesian Belief Net?
A: Bayesian nets are a network-based framework for representing and analyzing models involving uncertainty. They are used for intelligent decision aids, data fusion, 3-E feature recognition, intelligent diagnostic aids, automated free text understanding, and data mining.
They evolved from the cross fertilization of ideas between the artificial intelligence, decision analysis, and statistic communities. Bayesian nets have become useful for helping decision makers because of recent development of propagation algorithms followed by availability of easy to use commercial software.
Q: What is discrete event simulation?
A: Discrete event simulation is a branch of modeling and simulation which steps activity in a computer model each time an event is scheduled to occur, rather than at regular "delta t" time increments as in a continuous simulation model. This simulation approach allows modeling of complex decision situations over time where decision "events" can be triggered by activity in the environment and can change the behavior of components and the nature of consequent events.
Q: What is systems engineering?
A: Systems engineering is an interdisciplinary approach encompassing the entire technical effort to evolve and verify an integrated and total life-cycle balanced set of system, people, and process solutions that satisfy customer needs. Systems engineering is the integrating mechanism across the technical efforts related to the development, manufacturing, verification, deployment, operations, support, disposal of, and user training for systems and their life cycle processes. System engineering develops technical information to support the program management decision-making process. For example, systems engineers manage and control the definition and management of the system configuration and the translation of the system definition into work breakdown structures.
Adapted from EIA/IS 632, Processes for Engineering a System