|
A Basic Approach in Decision Analysis
A basic simplified approach in decision analysis involves the following four steps:
- Given a particular problem situation, the analyst develops a list of decision alternatives that will fix the problem. The decision maker is to consider these alternatives. A safety manager, for example, may be concerned with choosing a particular form of mitigation(s) to address specific hazards within a system risk. Consequently, the problem to be considered is the system risk, and there may be many alternatives to reduce this risk to an acceptable level, since there may be many cut sets of hazards and controls to evaluate. Each cut set is a particular path towards a pay-off.
- The second step involves determining and listing all possible future events associated with the problem situation — the system risk. These future events, which may not be under the control of the decision maker, are referred to as the "states of nature" for the problem. It is assumed that the list of possible states of nature includes everything that can happen and that the individual states of nature do not overlap. The states of nature are defined so that one and only one will occur. The alternative states of nature may refer to particular hazards and/or hazard controls that are appropriate to reduce the system risk associated with a particular cut set.
- Once alternatives and states of nature are defined, it is now possible to determine outcomes or what is referred to as "pay-off." Pay-off is a measure of the outcome, such as cost, loss or savings. Each combination of a decision alternative and a state of nature has a specific pay-off. Success paths may address mitigation, and failure paths may address hazards. Success or failure decision models can be used.
- The last step is to develop a pay-off table to present the information developed.
Table 1 — An Example Pay-off Table.
Click to enlarge
Using the pay-off table, the decision maker may base a decision on the following criteria:
Maximin (or Minimax) — Maximin is a pessimistic or conservative approach to reaching a decision. The decision maker attempts to maximize the minimum possible profits. Using the information contained in the pay-off table, the decision maker would first list the minimum pay-off that is possible for each decision alternative. The decision maker would then select the decision from the new list that results in maximum pay-off.
Maximax (or Minimin) — In using the criteria for maximization problems, the decision maker selects the decision that maximizes expected maximum pay-off. The decision maker first determines the maximum pay-off possible for each alternative. The analyst then identifies the choice that provides the overall maximum pay-off.
Interpretation of Pay-off Table
From a safety management point of view, determining pay-off in the above example is based upon projected losses and estimated loss prevention as a result of the implementation of the proposed risk control. Experience is required to make an appropriate judgment. An apparent choice would be to select the implementation of the administrative control, since it expends the least resources for the most pay-off. However, consider that the information presented may be based upon the most favorable assumption, in that it is expected that the administrative control will reduce the risk. It may be more prudent to expend the resources to eliminate the risk by re-engineering. In this case, the associated assumptions may be sounder. It is known that re-engineering the system would totally eliminate the problem (risk). It is important to fully define and understand assumptions supporting the analysis, and not to totally rely on the analysis presentation.
« previous page | next page »
|
