In this assignment, you will write a simulation of a train unloading dock.  Trains arrive at the          station as a Poisson process on average once every 10 hours.  Each train takes between 3.5 and 4.5 hours, uniformly at random, to unload.  If the loading dock is busy, then trains wait in a first-come, first-served queue outside the loading dock for the currently unloading train to finish. Negligible   time passes between the departure of one train, and the entry of the next train (if any) into the         loading dock---unless the entering train has no crew (see below).

There are a number of complications.  Each train has a crew that, by union regulations, cannot work more than 12 hours at a time.  When a train arrives at the station, the crew’s remaining work time is uniformly distributed at random between 6 and 11 hours.  When a crew abandons their train at the    end of their shift, they are said to have “hogged out” .  A train whose crew has hogged out cannot be moved, and so if a hogged-out train is at the front of the queue and the train in front finishes              unloading, it cannot be moved into the loading dock until a replacement crew arrives (crews from     other trains cannot be used).  Furthermore, a train that is already in the loading dock cannot be          unloaded in the absence of its crew, so once the crew hogs out, unloading must stop temporarily       until the next crew arrives for that train.  This means that the unloading dock can be lie unused even if there is a train in it, and even if it is empty and a (hogged-out) train is at the front of the queue; in  both of these cases we say the loading dock is “hogged-out” . More specifically, the loading dock is  hogged out if there are any trains in the queue but the loading dock is not actively unloading. (A bit  of thought should convince you that trains in the queue behind the one at the front do not affect the   hogged-out status of the unloading dock.) In other words, the loading dock can only be called “idle” if no trains are present.

Once a train’s crew has hogged out, the arrival of a replacement crew takes between 2.5 and 3.5 hours, uniformly at random.  However, due to union regulations, the new crew’s 12-hour clock starts ticking as soon as they are called in for replacement (i.e., at the instant the previous crew  hogged out); i.e., their 2.5-3.5 hour travel time counts as part of their 12-hour shift.

You will simulate this system for 1 million (1,000,000) hours (plus the time it takes for all remaining trains to depart), and output the following statistics at the end:

1.   Total number of trains served.

2.   The population average and maximum of the time-in-system over trains.

3.   The percentage of time the loading dock spent busy, idle, and hogged-out. Do these add to 100%?  Why or why not?

4.   Time average and maximum number of trains in the queue.

5.   A histogram of the number of trains that hogged out 0, 1, 2, etc times.

Input Specification: We will run your code, but to make it easy for the grader, we need everybody to adhere to the following guidelines: create a makefile that builds your program (if necessary), and your program should take either two or three command-line arguments. When given three                arguments, the first argument should just be “-s”, your program should read the second argument as thefile path to the pre-determined train arrival schedule (described below), and it should read the   third argument as thefile path to the pre-determined travel-timesfor new crews (also described       below). When only given two arguments, the first argument must be the average train inter-arrival time, and the second argument must be the total simulation time.  Thus, for example, if your program is written in C and compiled to an executable called “train”, then to run it with the default parameters above, I should be able to run it on my Unix command line as:

$ make

$ ./train 10.0 1e6 # 1e6 = 1 million hours

or

$ ./train –s schedule.txt traveltimes.txt

If you are using a language that does not run like the above (e.g. Python “python train.py 10.0 1e6”, or Java “java -cp . train 10.0 1e6”) create a shell script or program wrapper that takes in the arguments and runs your code as above. For example, if you are using Python, you can create a shell script named “train” containing:

#!/usr/bin/env bash

python train.py “$@”

That will allow the grader to run your program using the same syntax as above:

$ ./train 10.0 1000000

(Note: If you want to use this yourself on GNU/Linux, you’ll need to mark it as executable using the command “chmod –x train”)

Also, if you are using a language that does not require building/compiling (e.g. Python), just create a makefile with no targets:

target: ;

The pre-determined train arrival schedule contains three space-delimited columns: arrival times, unloading times, and remaining crew hours (in that order), with each arrival event on a new line:

0.02013 3.70 8.92

8.12 4.12 10.10

12.52 3.98 7.82

...

1210.0 4.12 9.21

The pre-determined travel-times for new crews contains a single column of data: the travel-timefor new crews. It would be safe to assume there could be more rows in this file than the previous file.

2.51

3.0001

...

2.89

When using a pre-determined schedule and there are no more train arrivals scheduled, end the    simulation after the very last train has departed; when using random values, stop adding arrival  events after the total simulation time has passed as specified by the command arguments (e.g., at 1,000,000 hours), but don’t stop the simulation the last train has departed.

Output Specification: Your program should print one line for every event that gets called.  We    want to be able to follow what’s happening in your code.  Each train and each crew should be        assigned an incrementing integer ID . The final statistics should come after the simulation output   and closely match the specified format. Output lines should resemble the following example (lines have been split here just for human readability but shouldn’t be in your output):

Time 10.03: train 0 arrival for 4.11h of unloading, crew 0 with 9.81h before hogout (Q=0) Time 10.03: train 0 entering dock for 4.11h of unloading, crew 0 with 9.81h before hogout Time 14.14: train 0 departing (Q=0)

Time 26.13: train 1 arrival for 4.24h of unloading, crew 1 with 7.49h before hogout (Q=0) Time 26.13: train 1 entering dock for 4.24h of unloading, crew 1 with 7.49h before hogout Time 28.94: train 2 arrival for 4.42h of unloading, crew 2 with 6.96h before hogout (Q=0) Time 30.37: train 1 departing (Q=1)

Time 30.37: train 2 entering dock for 4.42h of unloading, crew 2 with 5.54h before hogout Time 34.79: train 2 departing (Q=0)

…

Time 58.34: train 7 entering dock for 4.11h of unloading, crew 7 with 0.12h before hogout Time 58.46: train 7 crew 7 hogged out during service (SERVER HOGGED)

Time 61.81: train 7 replacement crew 8 arrives (SERVER UNHOGGED)

Time 65.80: train 7 departing (Q=0)

…

Time 721.40: train 58 crew 63 hasn't arrived yet, cannot enter dock (SERVER HOGGED)

…

Time 7201.55: simulation ended

Statistics

Total number of trains served: 721

Average time-in-system per train: 6.37h

Maximum time-in-system per train: 25.1h

Dock idle percentage: 59.95%

Dock busy percentage: 40.05%

Dock hogged-out percentage: 3.71%

Time average of trains in queue: 0.205

Maximum number of trains in queue: 4

Histogram of hogout count per train:

[0]: 594

[1]: 122

[2]: 5

... (show as many bins as necessary)

Numbers can have any number of decimal places (valid examples include 13 and 3. 14159), but no scientific notation.

A script to automatically check the output formatting is provided to give feedback on I/O           specification compliance. Look on openlab in the directory /home/wayne/pub/cs115/A1-syntax- checker. To test your program’s input and output, you can run the following commands:

$ cd /home/wayne/pub/cs115/A1-syntax-checker

$ YourTrainDir/train –s schedule.txt traveltimes.txt | # <- pipe ./train_format_checker.py

If your output violates the specification as checked by the above checker so   badly that we can’t test your output for correctness, we reserve the right to give you a zero.

Submission: Submit a brief write-up of your results, along with your source code and specific       sections of the output (only the first few pages and the last page) for the default parameters.  There should be no more than 10 pages. Submit both a paper printout to me in class, and also electronically via the submit command on ICS openlab.