Monte Carlo Simulation
Monte Carlo application is accessible through the Launchpad:
After clicking on Monte Carlo tile, you are in Monte Carlo Simulation application as follows:
Inputs: This is a table of input cells with corresponding distributions and parameters, and the target cell. The structure of the table is like this:
| name1 | Cell 1 | distribution 1 | parameter 1 | parameter 2 | … |
| name2 | Cell 2 | distribution 2 | parameter1 | parameter2 | … |
| … | … | … | … | … | … |
| … | Cell k | distribution k | parameter 1 | parameter 2 | … |
| … | Cell m |
Where name1, name2, … are the names will be used in the repost for corresponding cells. Cell 1, Cell 2, … are the location (address) of the input cells and Cell m is the location (address) of target cell.
The available distributions with available parameters are:
| norm | mean | sd | ||
| triangle | min | mode(likeliest) | max | |
| unified | min | max | ||
| lnormal (log normal) | meanlog | sdlog | ||
| beta | min | max | shape1 | shape2 |
| cbpert (classic beta pert) | min | likeliest | max | |
| vbpert (vose beta pert) | min | likeliest | max | |
| gamma | min | shape | scale | |
| weibull | min | shame | scale | |
| maxExtrem (evd) | likeliest | scale | ||
| minExtrem | likeliest | scale | ||
| logistic | mean | scale | ||
| student t | median | df | scale | |
| exponential | rate | |||
| pareto | min | shape | ||
| binomial | trials | probability | ||
| poisson | lambda | |||
| hypergeometric | success | trials | population | |
| nbinomial (negative binomial) | trials | probability | ||
| geometric | probability | |||
| discrete uniform | min | max | ||
| true-false (yes or no) | size | probability of true | ||
| csample (continuous sample) | obs 1 | obs 2 | obs 3 | … |
| dsample (discrete sample) | obs1 | obs 2 | obs 3 | … |
For instance, we have three inputs located in Cells B1 to B3 and the result (target cell) is in B4. Also, imagine that for inputs from B1 to B3, we have a scenario with normal distribution with mean of 5.5 and standard deviation of 1.5, a scenario with beta distribution of minimum 3, maximum 7, shape 1 equal to 1 and shape 2 equal to 1.2, and lastly you have a sample of discrete numbers of [1, 3, 7, 10, 12, 15, 20] for the last scenario, your scenario table looks like this:
| B1 | norm | 5.5 | 1.2 | |||||
| B2 | beta | 3 | 7 | 1 | 1.2 | |||
| B3 | csample | 1 | 3 | 7 | 10 | 12 | 15 | 20 |
| B4 |
Correlations: This is matrix of correlation between your scenarios. Default is 0 correlation. For instance, in our previous example if B1 (norm) has 0.45 correlation with B2 (beta), and B2 (beta) has -0.36 correlation with B3 (csample), your correlation matrix would be like this:
| 1 | ||
| 0.45 | 1 | |
| 0 | -0.36 | 1 |
You just need to provide the lower triangle of your correlation matrix. Anything in upper-triangle will be ignored by the function.
method: You have the option of Simple Random Sampling (SRS) or Latin Hypercube Sampling (LHS). If any empirical sample is provided as part of distributions, the function sets method to “LHS”, automatically.
Generate Report: If set True a comprehensive report is generated by the function. It is False by default.
Run Simulation: You can define the number of simulations. This is the number of scenarios which the user would like to generate. If n is less than 5, the function runs with n = 5 and if n is greater than 10`000, the function sets n = 10000. By pressing this key, the application generates the output and report them in a separate tab.
Example:
Here is an example of a report generated by this application. As you can see, an interactive report is generated in the task pane which you can expand in your browser:
