E-mail correspondance regarding SensIt
Here is an Excel file that computes the posterior probabilities for the Sunspotz problem in light of experimentation results.
No experimentation: Decision tree for the Sunspotz problem without experimentation. Note that you should not enter any data in cells which have formulas.
With experimentation: Decision tree for the Sunspotz problem with experimentation. Note that you should not enter any data in cells which have formulas. Also note here that the posterior probabilities are calculated automatically from the table to the right of the tree.
Here are two simple finance examples where the input variables are the present value (P), interest rate/year (i), and the number of years, and the output variable is the future value (F).
In the Tornado diagram, we fix all inputs at their base values and find the range of values the output would assume.
Sensitivity analysis with SensIt: In this file we use SensIt and calculate the new maximum EMV values for varying values of the probability of high demand which is varied from 0.1 to 0.5.
We need utility theory guide us where EMV fails. Consider the case of a Deal or No Deal contestant who turns down a certain $100,000 for a small chance of winning $300,000 and a large chance of getting nothing. This person is mosty likely a risk-seeker. We buy insurance even though the expected loss is smaller than the premium. Are we then risk-averse?
In this file we assume that Sunspotz is risk-averse. With the utility function u(x) = 10*((100/6.1)*(x+1.1))^(0.5) shown at the bottom of the file the optimal decision is not to do the experimentation. Note here that x ranges from -1.1 to 5, and so u(-1.1) = 0 and u(5) = 100.
Risk aversion and related concepts (e.g., risk premium) are important in decision analysis. I am providing only a brief discussion of these in the lecture, but I would like to ask the students (especially, the PhD students) to read carefully my notes in this link.
There is also a nice Wikipedia article in this link that is worth looking at.
2011-11-16: Here's the e-mail from the book rep:
`` Hello Professor Parlar, My apologies that it took so long to get you this information; I was given quite the run-around from our US counterparts trying to get this information for you. Below is the answer that I received from the US marketing managers regarding your question: "The latest version (that works with Excel 2010 and 2011) can be downloaded at treeplan.com. It is a 30-day trial version, but seems to continue working after that; if I remember correctly this version can be used on the same terms as the textbook version (i.e., you can use for the duration of the course, but if you use after that, you should pay the license fee). We will have an alert included on the book's website which reads as below: The version of TreePlan that was packaged with the book does not always operate properly with Excel 2010. However, the latest version works with both Excel 2010 and 2011 (as well as earlier versions). A 30-day trial of the latest version can be downloaded at treeplan.com." I hope this answers your question regarding treeplan. Kind regards, Shannon''
2011-11-02: In this chapter we will use the Excel add-in TreePlan. The free version you have received works with Excel 2007, but I am told that it has problems with Excel 2010.
I have asked the publisher if we could have access to the 2010 version. The response on 2011-11-02 was `` Hello Professor Parlar, I am going to look into this and will get back to you as soon as possible. Kind regards, Shannon.''
In the meantime, you may be interested in downloading the free trial version for 2010 from here. This version will work for 30 days only. Since the final exam is on December 13, 2011, make sure you install it after November 14, 2011.