HELP: Discrete Event Model and Simulation

Post models or parts of models you would like to share with other ExtendSim users here.
Post Reply
extendnovice
Posts: 7
Joined: Sun Jul 21, 2019 4:45 pm
First Name: Paul
Last Name: Davis

HELP: Discrete Event Model and Simulation

Post by extendnovice » Tue Aug 13, 2019 10:25 am

I am new to modeling and don't know where to start in ExtendSim. PLEASE HELP.

Create a simple, discrete event model and simulation to represent a system responding to calls for fires. Explore the implications regarding system design and develop insights from analysis of model output.

We propose use of rail guns to respond to calls for fires. We could as easily propose to examine directed energy weapons used against multiple targets where there are re-charging, targeting and dwell times associated with each engagement. Don’t lose the forest for the trees by concentrating too hard on a specific weapon type.

You are comparing several different architectures for providing fires support to an expeditionary force. The architectures involve the use of one or more rail guns. The cycle time (to prepare, charge, and fire each gun) is a random variable.

Targets appear randomly and result in generation of fire requests on average every minute. The time to prepare, charge, and fire each gun has a mean of twenty seconds. Assume, initially, that each fire request is considered only once in the operational cycle – so there is no second cycle for any particular fire request.

Consider the period for analysis to be 3 hours (180 minutes).

1. Construct an ExtendSim discrete event model for a single rail gun system, using the exponential distribution for all arrival and server times. Use ExtendSim’s notebook feature for monitoring parameter and output values. Use an ExtendSim database to read the target and gun parameters into the model (note req’s 8, 10), and to record your output metrics for each run.

2. Estimate the average time until a request is answered – that is, the time to fulfill a fire request (exclusive of the rounds’ time of flight – which is irrelevant for this analysis). Provide a 90% confidence interval for your answer.

3. Estimate the fraction of the average time in (2) attributable to queuing delays.

4. What does it mean for a system to have transient behavior? Is there a transient (perhaps a warm-up) state for your model?
If so - What metric(s) are you using to identify the transient state? If a transient state exists, will you condition your output data to account for the transient behavior – if so, how? - if not, why not?
If not – what rationale are you using to support the presumption of no transient?

5. On average, how many fire requests are waiting for a gun to become available? Provide and interpret an approximate 90% confidence interval for your answer.

6. Identify and examine the maximum time to wait (queueing time) and the maximum time to respond (total time in the system) to fires requests. Why would these performance characteristics be of interest to a system designer or to an operational commander?

7. As an excursion, when supported forces are in close contact with the enemy, the mean time between fires requests can drop to twenty seconds (0.333 minute) – that is, requests for fires occur more frequently. Does that change your answers to (2) to (5)? How so?

8. Implement one refinement to the model of this system. How does your refinement affect your answers to (2) through (6)? Note, a refinement is not a simple parameter change (such as, the cycle time was 1.0 and now it’s 0.5) but rather an improved model that more closely emulates the system being studied.

9. Use the un-refined model (model prior to Req 9) and compare alternatives. Is it preferable to have:
- one rail gun with a mean cycle time of 20 seconds (baseline configuration) or,
- two rail guns, each with a mean cycle time of 40 seconds or,
- four rail guns, each with cycle times of 80 seconds (1.333 minutes)?
When the mean inter-arrival time of requests are:
a) 1 minute
b) 20 seconds
Why? Support your assertion - statistical hypothesis tests may be appropriate for your arguments.

10. This model may be sensitive to the choice of fire request and server distributions. Use your baseline model to explore this notion by setting the time between target appearances to follow a normal distribution with mean 1.0 and standard deviation of 0.2. Set the gun cycle times to follow a normal distribution with mean 0.333 and standard deviation of 0.1. Do the answers for (2) to (5) and (7) change? Explain your results using common sense arguments.

ceciledamiron
Posts: 55
Joined: Wed Jul 02, 2008 12:10 pm
First Name: Cecile
Last Name: Damiron

Re: HELP: Discrete Event Model and Simulation

Post by ceciledamiron » Tue Aug 13, 2019 11:36 am

Hello David,

It looks like you are posting your homework assignment.
It would be unfair to the other students to help you. Also your teacher expects you to figure things by yourself.

We have a Discrete Event Quick Start Guide which you can use as a starting point. It will give you a good start in how to use our tool.

Good luck with your assignment, I hope you will have a good experience with ExtendSim.

Cecile Pieper

extendnovice
Posts: 7
Joined: Sun Jul 21, 2019 4:45 pm
First Name: Paul
Last Name: Davis

Re: HELP: Discrete Event Model and Simulation

Post by extendnovice » Tue Aug 13, 2019 1:22 pm

Yea sorry didn't mean to make it seem like I wanted someone to do it for me. Just an idea of what all I have to do. I THINK I have a starting point. I'll use the Discrete Event Quick Start Guide and see if I can make some progress.

Post Reply