Decision analysis is a process whereby a description of a process is transformed into a quantitative model that can be used to demonstrate the relationships between inputs used, processes and outputs. This facilitates the efficient application of analytical methods

^{1} so as to determine what can improve a process, sensitivities of a process performance to specific inputs and conditions and to project the implementation of a project over time or the performance of a process in the future.

To achieve useful practical outcomes from such a process there is a need for a proficiency in the design and implementation of appropriate decision analysis models. The most useful models often combine well-established analytical techniques

^{1} with the unique features, relationships, algorithms and structural logic relating to the specific system of concern to decision-makers. Best practice in the identification of the most appropriate decision analysis models involves a reiterative decision analysis cycle.

The decision analysis cycleModel building has the purpose of maximising the transparency of all critical factor relationships which determine decision outcomes while minimising the potential for errors arising from inadequate model design and information used. The decision analysis cycle is a design refinement and learning approach based upon a proactive evolution in the quality of three specific types of information:

1. The information and knowledge used to construct a model which encapsulates the understanding of the relationships between the critical factors which determine decision outcomes (

*deterministic*). Examples are the number and quality of components required to make a device or the amount of fertilizer required to achieve a specific crop yield.

2. The confidence in the applicability of the model rests upon the probabilities of critical events, not all of them under the control of the decision-maker (

*probabilistic*). Examples include variability in output of a crop associated with weather, irrigation and other protective devices or the likelihood of frost attack.

3. The utility of the model finally depends upon the availability and quality of the information (data) used in building the model and to predict outcomes (

*informational*). Examples include more refined data on the relationships between determinants (item 1), better statistical information on probabilities (item 2) or new information on alternative determinant relationships (options).

A decision analysis model is deemed to represent an acceptable standard when sufficient confidence has been attained that the deterministic, probabilistic and informational dimensions cannot be refined further have attained an acceptable level of uncertainty

^{2}.

This site is sponsored by SEEL Systems Engineering Economics Labs http://www.seel-telesis.com