Decision Analysis

Decision analysis is a structured method for making optimum decisions under uncertain settings. It is utilized by organizational managers faced with various possibilities. Decision tree analysis is the oldest and most prevalent type of decision analysis.

Decision analysis is a technique that aims to give a rational foundation for management choices made under ambiguity. Decision analysis is a methodical, statistical, and expansive approach to making complex judgments. It is a normative technique for choosing between acts with unknown results.

• Decision analysis is a methodical, statistical, and expansive approach that aims to give a rational foundation for management choices made under ambiguity.
• Decision analysis is a normative technique that helps arrive at complex judgments while choosing between acts with uncertain results.
• It heavily relies on the quality and completeness of model inputs and the analysts' assumptions.
• The decision tree is a technique of decision analysis. It comprises considering decisions, their chances in terms of possibility, and their consequences in terms of further decisions and their outcomes.

Decision analysis is considered to be interdisciplinarity. However, traditionally it is considered to be a subfield of operations research. Thus, after academic advancements, it has matured into a professional field. Since the late 1950s, the technique has been utilized to support commercial and public policy decision-making. It has been widely used in the pharmaceutical, oil, and gas sectors because of the frequent necessity of making high-stakes, high-impact choices.

The ability of decision analysis modeling to calculate a range of probable values around a given mean is a significant strength. This method is known as "sensitivity analysis." Hence, it enables the user to better comprehend the likelihood of making a poor decision if a strategy is implemented.

Also, it is worth noting that decision analysis is heavily dependent on the quality and completeness of model inputs and the analysts' assumptions. For instance, it is possible for drugs to have unanticipated adverse effects or for treatments to have long-term costs that are not readily evident to analysts. Any of these can result in unsatisfactory consequences.

In brief, decision analysis is a normative technique for choosing between acts with unknown results. Such outcome uncertainty can be represented by a probability distribution for variables representing the most significant outcomes of the acts under consideration. Hence, a utility function characterizes the relative preference of the decision maker for the many possible outcomes. It incorporates the decision maker's risk aversion. Then, a logical decision maker should choose the action that maximizes a certain mathematical combination of the determined probabilities and utilities.

Decision analysis is a technique that aims to give a logical framework for making one management choice out of several options under ambiguity. Moreover, this approach is based on a structure known as a decision tree.

Mainly, a decision tree has two sorts of nodes: choice nodes and chance nodes. A choice node is a node where a decision must be taken, whereas a chance node is a node where an unexpected consequence is achieved. Hence, movement from node to node in the tree indicates the passage of time. Therefore travel from node to node represents either the need to make a decision or the realization of a random conclusion.

A decision-analytic model, often a decision tree, is the core instrument of decision analysis. It gives a graphical representation of the sequences of events that might occur following various options (or acts). It also considers the consequences associated with each pathway.

Choice models can integrate the probability of the underlying (actual) states of nature when calculating the distribution of potential outcomes for a specific decision. The decisionmaker is unaware of these possibilities, yet they are very essential.

The term "model" has different connotations in various contexts. Similarly, the defining characteristic of a decision model is that its purpose is to facilitate decision-making and not to establish truth claims.

Evidently, the objective of statistical research designs is to collect data. However, the objective of decision analytic studies is to analyze evidence, albeit, in the absence of a systematic and formal strategy that supports the decision maker in processing the (sometimes divergent and complicated) data, the processing is conducted less formally.

Let us look at the following decision analysis examples to understand the concept better.

Example #1

A decision analyst is asked to consider and evaluate the option of installing a new machine in the production department; hence to come to a decision, the analyst decides to use the decision analysis tree technique.

Moreover, the decision analyst knows that the value of installing a new machine depends on the chance that the capacity will require expansion in the future. Accordingly, it influences the probabilities of true and false positive results and true and false negative findings. The relative values attributed to these various outcomes establish the usefulness of a machine. For instance, how detrimental is it to fail to install new machinery without increasing capacity relative to not installing a machine and having the need to expand in the future?

Also, the analyst considers if the knowledge acquired from the decision tree alters the earlier choice that would be taken. Likewise, does this alteration results in better future outcomes is also considered. Particularly in circumstances where various data inputs from a range of research are relevant to a specific decision-making context, decision analysis approaches have been proven to be very beneficial.

Example #2

Assume you are the Financial Manager of Car Plaza, a globally renowned upscale car service station. Your station has been performing well, and hence you intend to capitalize on this by launching a new location. Additionally, you are considering choosing between Ontario and British Columbia.

Also, your team has provided you with a range of quantitative analyses for review. So, as an analyst, you will have to make a choice using available information. Your team determined that the weather significantly influenced the station's profitability for each city. However, these places show a big contrast between winter and summer. Like Ontario gets snowstorms, and British Columbia gets abundant rainfall. Moreover, your team informed you that, in both instances, when the weather is unfavorable, individuals prefer to avoid visiting the car station.

Below are the net present values (i.e., a profitability metric that represents payoff) and probabilities of various weather patterns for each choice. You must choose the city with the highest return while considering the weather risk. So following are the variables and probability of it that you consider. Hence, the final decision of choosing a location will be based on that.

Ontario Alternative:

British Columbia Alternative:

Solution:

The weighted-average payoff for each alternative.

Ontario:

=35,00,000*0.4+20,00,000*0.3+15,00,000*0.3

=14,00,000 + 60,00,000 + 4,50,000

=$24,50,000.

British Columbia:

=35,00,000*0.2+15,00,000*0.4+500,000*0.4

=70,00,000 + 60,00,000 + 2,00,000

=$15,00,000

The best alternative is thus Ontario.

This has been a guide to Decision Analysis and its definition. We explain the techniques of decision analysis, decision tree, along with examples. You may learn more from the following articles -

• PESTLE Analysis
• Statistical Analysis
• Pareto Analysis