Getting Started

Rational Will is a multi-criteria decision analysis software that will help you to make rational decisions.

Students of Decision Analysis course will find this tool extremely useful for learning various concepts from Normative Decision Theories, Utility function from Behavioral Economics, Sensitivity Analysis, Monte Carlo Simulation, Stochastic Dominance, Multi-criteria decision analysis and lots more. Even Markov Chain and Markov Decision Process modeling are included within Rational Will. Rational Will is intuitive enough with the modern user interface and user experience that it can be used easily in daily life for solving real-life problems. When you start the application, you will be presented with the following screen.

main_screen

From the above screen, you can choose the modeling tool that you want to use.

The following modeling tools are available in Rational Will.

1. Decision Tree
2. Decision Matrix
3. Analytic Hierarchy Process
4. Influence Diagram
5. Markov Decision Process
6. Pros and Cons Analysis.

The In order to use the Analytic Hierarchy Process, please choose the Decision Matrix tool, because, Analytic Hierarchy Process can be modeled using the Decision Matrix modeling technique.

Every modeling tool has it's own getting started page in this documentation. So, please navigate to the corresponding getting started pages for learning details about the specific modeling tool described in this section.

When to choose what modeling tool?

Decision Tree

Your decision problem may not be limited to a set of option selection, rather your options can lead to some further decision scenario along with uncertainties. For example, notice the following scenario where you want to decide, should you go by plane or car. If you go by car, the car can break down and you may need to decide what you will do if that happens. A decision tree is what you need to model such a scenario. You can always have your Plan-B ready with your decision tree.

decision_tree

The Payoff of a node can be modeled with multi-criteria objectives, advanced utility functions, probability distributions etc.

Decision Matrix

In the decision matrix, you will start with your objectives. It is not rational to care about every pro and con of an option. Rather, identify your objectives and your options and then evaluate the options based on your objectives. Notice the following decision problem. Say you want to buy a car and you have identified 3 options. You have also identified 4 objectives "Must have Android Auto", "Minimize Price", "Maximize Fuel efficiency", "Must have GPS Navigation". So, based on these 4 criteria, you can use the Decision Matrix to model your decision problem as shown here.

car_selection

Rational Will Decision Matrix is an advanced decision analysis tool, which is not limited to model just the decision matrix. Rather, you can model uncertainty for every criterion of an option. You can even incorporate Probability Distribution. Then Monte Carlo Simulation can be generated. The risk profile is generated for any uncertain situation. And lots of metrics are calculated, along with sensitivity analysis. The benefit to Cost ratio, Minimax regret analysis etc. are just a few examples of what can be done with Decision Matrix. You can model an advanced utility function in Decision matrix too.

Analytic Hierarchy Process

When you want to select from a set of identified options and when you have identified your objectives, then Decision Matrix is the right tool. Decision Matrix support a variety of objective attribute types, like Subjective, Boolean, Number, Monetary, Categorical etc. If all of your objectives are a Subjective type, then Decision Matrix present you Analytic Hierarchy Process modeling User Interface. Analytic Hierarchy Process is a particularly useful method when the decision maker is unable to construct a utility function.

analytic_hierarchy_process

Markov Decision Process

When you have a set of situations that can be repeated based on some set of your actions, then you may want to evaluate a policy. The policy tells you, what action to perform in what situation. For example, you run a store and you have 3 situations, "Good sale", "Bad sale", "Moderate Sale". In these states, you can take some actions like "Advertise more", "Advertise less". Now, you notices that, when you are in Good sale state, more advertising just costs money but does not increase sale. But if you do not advertise more, you may lose the sale. So, you may need to find a policy, should you advertise more when you have a bad sale? should you advertise less when you have a bad sale? etc. Such scenario can be modeled by a Markov Decision Process.

Rational Will Markov Decision Process analyzer will calculate a policy for you and then, based on recommended action, it will generate a Markov Chain for you.

Markov_ Decision_ Graph

Influence Diagram

When you have a big problem to solve but you do not want to start the analysis in deep, rather, you want to jot down the problem in high level, you should use the Influence Diagram.

Influence_diagram

Pros and Cons

If you are in a dilemma about anything like, "should I do it or not", "should I get a divorce?", like this, and you don't want to spend much effort on deep analysis, then a Pros and Cons analysis can serve the purpose quickly.

Should_ I_get_a_divorce

Even though, a Pros and Cons are not recommended for a better decision. Always try to use Objective based decision analysis. Therefore, the Decision Matrix should be the next modeling option to choose.

Last updated on 23 September 2018, Sunday, 5:20:49 PM
If you have any questions or concerns about this tutorial, Please feel free to share your comment.