Part A:
1- Considering the same parameters for all stages of the process (slides 30 to 33) and the same duration of the simulation replica (20 days, working 24x7h), make a comparison on the following indicators between a model created in FlexSim and another in Arena:
1- Average duration of each request in the modeled process
2- Average cost of reviewing a request
3- Maximum review time for a request
4- Maximum number of applications awaiting review
5- Percentage of occupation of the risk manager
Part B:
2- Implement the modifications discussed in following 2 Exerciseds with Arena or FlexSim, and comment on how this affects the maximum number of requests that come to be under review.
Exercise 1: Add a new filter or control task (process in the tool) prior to the review of the mortgage application • The filtering time is a minimum of 20 minutes (m), with 25m being normal, and a maximum of 50m • Assign a “Receptionist” to carry out the filtering. The cost per hour of this worker is €6.75/hour
Exercise 2: Return some requests after the filtering process of the previous exercise • 8% of the requests do not pass the filter • Now the percentage of requests that, once filtered, are approved goes from 88% to 94% and the review time is reduced by 10% • How do these changes affect the cost of the process? And the total review time of an application? Try to respond to the same points as in the previous exercise: percentage of occupancy of the risk manager, number of requests waiting, etc.
3- The formula to obtain the cycle time of a section of a process in which there is a rework, that is, a part of the process in which with probability r, a set of tasks with a duration of time T would have to be performed again. (see figure), is given by the expression T1−r.
Build a new model, in Arena or FlexSim, that allows you to verify that, indeed, the execution times of a rework are close to the value obtained after applying the formula. You can freely choose the processing and arrival times between entities. The model has to be very simple and consist only of a process with a task that can be iterated again with probability r.
Ordinarily, why would times be different in reality? Think about the justifications that exist for using models and simulation techniques, and what parameters of those used in the example models on the transparencies, you would not specify in the requested model so that the times of the formula and those obtained after the executions coincide ( or were very similar).
In your model, and in light of the simulation results obtained in your model and with the time parameters that you have established, would you be able to determine how many resources (units) would be necessary so that the times of the formula and those of a model that used resources Would they also coincide?