Decision analysis using Hiview3
Hiview3 models are built to aid decision makers in choosing between mutually exclusive options. Many are built to choose between potential investment plans. These investment plans have included market expansion plans for hair care products, choosing a partner for product development and siting a nuclear waste facility. There are 5 main stages to modelling in Hiview3.Stage 1 - Constructing a model
The first stage in Hiview3 modelling is to construct the outline of the model. On the model structure above, each node on the bottom level of the tree is a criterion against which investment options will be measured. They have been grouped under nodes of similar criteria to clarify the model. This becomes more important when setting preferences. The topmost node represents the whole decision to be made. It is usual for the two nodes directly beneath to represent costs and benefits.
Each of the investment/action options in the model are now scored against the criteria set out in the tree structure. Not all criteria are monetary. Part of the power of Hiview3 and MCDA is that it allows quantification of non-monetary assets. This is especially important for public sector users who do not have the luxury of monetary targets like profit. Scores can be input to a Hiview3 model either numerically or graphically.
This is the key differentiator of MCDA over other decision analysis techniques. At this stage the decision maker(s) must make a value judgement on the relative importance of different aspects of the model. Preferences are set using swing-weighting techniques. These are technically more robust than simply setting relative importance of criteria. Often, it is a group of senior decision makers who perform this task. Catalyze provides Decision Conferencing services to manage this process. It is however, entirely possible for one person to make decisions this way. Buying a new car is a good example of this.
When the model is populated, various analyses can be performed. Sensitivity analysis is used to demonstrate the robustness of an option over its rivals. It is easy to see graphically where one option is completely dominated by another. Mapping nodes against each other highlight which options are the most efficient at each level in the model.
Hiview3 is designed to help support more effective decision-making. It does not make the decision for you. However, it does provide a recommendation of the most appropriate option. The graphic above shows the results graph of the top node. This node represents the decision. The option with the highest weighted score is the recommendation.
To view an example Hiview3 report, click here.