Integrated Siting Systems Uneeb ul haq Mughal 16110030 Sahrish Jaleel Shaikh
16110299
Mehroze Munawar
16110197
Myra Iltefat
16110278
Integrated Siting Systems, Inc. has developed a satellite-based tracking system that continuously updates and displays the location of every vehicle in its network. A recently signed contract presents new technical challenges that may make the normally installed system inadequate. While it is possible to augment the system so that it will definitely work, doing so is costly. Alternatively, it is possible to conduct an imperfect test that will improve the company's information about whether the normal system will work. Current simulations indicate a 90% chance that the standard system will meet the performance requirements of the contract. Better estimates of the environmental parameters that determine this factor could be obtained and a new series of simulations could be run for $50,000. The System Design team has determined that a standard system will work 96% of the time if it were to pass this new test and, even if it failed there would be a 72% chance that it would still work anyway. Influence diagram (Exhibit 4) The two decision nodes (represented by the rectangles) pertain to the following two decisions 1) Whether the test should be conducted 2) Which kind (standard or robust) of system to install for the new project The two chance nodes (represented by the ovals) pertain to the following chance events 1) The results of the test 2) Whether the system turns out to be good or bad The diamond represents the final target i.e the profit/ payoff. The condition of the system and the decision of whether or not to implement the test will determine the test result. The test result in turn will influence ISSI’s decision of which type of system to install. The decision of the system type along with the actual condition of the system (good/bad) will finally determine what the final payoff will be.
Intangible Costs: Contract terms stipulate that ISSI’s obligations should the system fail include pre-specified penalties for a delay in system launch time and pulling and replacing the units with more
powerful processors and receivers. The complete and total incremental cost resulting from this in-field replacement project would be $400,000. However, this does not include any intangible costs of lost reputation resulting from our first major failure, negative media coverage, and any additional detracting factors. It is notable that if this cost of lost reputation can be quantified, then this will add to the $400,000 value mentioned above and the decision tree will be made accordingly then. The decision tree prepared for this analysis does not take into consideration this implicit cost. The decision tree (Exhibit 1) shows that two decisions have to be made in succession. The first decision involves whether or not to conduct the test. If the test is conducted, a cost of $ 50,000 is incurred. There can be two outcomes of the test if it is conducted:
The system passes the test (the probability for this as given by the case is 0.75).
The system fails the test (the probability for this as given by the case is 0.25).
Based on the test results, a second decision will be made pertaining to the type of system that will be installed (standard or robust).
If the test is failed and the robust system is installed the profit earned will be $150,000 (850,000-700,000).
If the test is failed and the standard system is installed the revenue earned will be equal to $300,000 (850,000-550,000). The overall profit however will depend on whether the system turns out to be good or bad. If the system is good, no cost will be incurred so the net profit will be equal to 250,000 (300,000-5000). The probability of this occurring (P(System good / Fails test)) is 72%.If the system is bad, an incremental cost of upgrading in the field from the standard to the roust system will be incurred. This cost equals $400,000 and so the net profit will be $-150,000 (300,000-5000-400,000). The probability of this occurring is 0.28 (P(System bad / Fails test)).
If the test is passed and the robust system is installed the profit earned will be equal to $150,000 (850,000-700,000).
If the test is passed and the standard system is installed the revenue earned will be equal to$300,000 (850,000-550,000). The overall profit, however will depend on whether the system turns out to be good or bad.. If the system is good, no cost will be incurred so the net profit will be equal to 250,000 (300,000-5000). The probability of this occurring (P(System good / Passes test)) is 0.96. If the system is bad, an incremental cost of upgrading in the field from the standard to the robust system will be incurred. This cost equals $400,000 and so the net profit will be $-150,000 (300,000 -5000400,000). The probability of this occurring is 0.04 (P(System bad / Passes test)).
Installation of the standard system generates higher payoffs whether the test passes or fails so that should be the option to go with in both cases, as shown by the decision tree. The EMV of conducting the test will be equal to $210,000. Even if the test is not conducted, either the standard system or the robust system will be installed.
If the robust system is installed the profit earned will be $150,000 (850,000-700,000).
If the standard system is installed the revenue earned will be $300,000. The profit earned will again depend on whether the system is good or bad. If the system is good, no cost will be incurred so the net profit will be equal to $250,000. The probability for this will be 0.9 (P(System Good)). If the system is bad, an incremental cost of $400,000 will cause a loss of $150,000. The probability for this will be 0.1 (P (System Bad)).
Installation of the standard system will generate higher payoff so that option will be selected in this case too. The EMV of not conducting the test will be equal to $260,000. The expected value of imperfect information in this case is the difference of the EMV value obtained by incorporating the probabilities associated with new information gathered by tests and the EMV value obtained without incorporating the information obtained from tests. According to the calculations, EVII is equal to 0 (=260,000-260,000). This value-of-information suggests that you should not spend anything to obtain this imperfect information because it’s
not worth the cost. In both the cases, the EMV remains the same i.e. information gathered does not influence the decision that you should make. The risk profiles in Exhibit 3 also support the decision of not taking the test. Probability chart in risk profile shows almost same dispersion in the payoffs but the decision of taking test has lower values (-150,000 and 250,000) as compared to the decision of not taking the test (100,000 and 350,000). Moreover in cumulative risk profile chart, as decision of taking test lies always to the left of not taking test and there is space in between them i.e. they do not perfectly overlap, we can say that Decision of taking test is stochastically dominated by decision of not taking the test. This result is important as it clearly rules out the option of taking the test and hence helps us to save the cost and time of designing the test. Optimum Strategy ISSI should not conduct the test and it should install the standard system for the new contract since these options have a higher EMV
Appendix___________________________________________________________________
Exhibit 1
Exhibit 2
Exhibit 3
Exhibit 4