cpp_training

Instructions

You need CMake to configure the project and the GCC or Clang compiler and associated tool chain.

1. Clone the project repo

git clone https://github.com/thinhong/cpp_training.git

2. Change current directory into the project folder, build an executable file using make

cd cpp_training/
make

3. Run the executable file

./DiscreteTimeModel

A message will appear asking you to set path to input file, output folder and output file name.

Enter path to input file: 
Set path to the folder you want to save output file (ex: /home/Documents): 

• Path can be absolute path or relative path
• The output folder will be created if it is not existed
• The output will be saved as output.csv in your given output folder
• If output.csv already exists in the output folder, it will be overwritten

Config file

{
    "simulationDuration": 10,
    "errorTolerance": 0.01,
    "timeStep": 1,
    "initialValues": {"S": 999, "I": 1, "R": 0, "V": 0},
    "parameters": {"beta": 1.2, "N": 1000},
    "transitions": {
        "S -> 0.3 * I": {"distribution": "mathExpression", "expression": "beta * S * I / N"}, 
        "S -> 0.7 * V": {"distribution": "constant", "constant": 2},
        "V -> R": {"distribution": "exponential", "rate": 1.5}, 
        "I -> R": {"distribution": "gamma", "scale": 2, "shape": 3}
    }
}

• simulationDuration: suppose the unit of time is days, this define the number of days for follow-up
• errorTolerance: the cdf of a distribution will never reach 1.0, this indicate when to round the cdf to 1.0
  • When the cdf reach 0.99 (which means errorTolerance = 0.01) it will be rounded to 1.0 and the day that cdf = 0.99 is number of subcompartments
  • If errorTolerance = 0 and the program run endlessly because it produces infinite subcompartments
• timeStep: if a day is binned into 100 smaller time bins, a time step will be 0.01 day, define here "timeStep": 0.01
• initialValues: number of individuals in a compartment at the first time step
• parameters: if we define a compartment transition using math expression, define all the parameters here
• transitions:
  • S -> I: compartment S transitions to compartment I
  • S -> I then S -> V: compartment S transitions to compartment I, then the rest of S transitions to I
• To define multinomial
  • 0.3 * S -> I, 0.7 * S -> V: S concurrently transitions to I and V, of which 30% moves to I and 70% moves to V
• distribution: distribution of waiting time, could be one of the followings
  • gamma (gamma distribution): parameters are scale and shape
  • weibull (weibull distribution): parameters are scale and shape
  • exponential (exponential distribution): parameter is rate
  • lognormal (log-normal distribution): parameters are mu and sigma
  • mathExpression (math expression): put any math expression to the mathExpression parameter, the expression can contains compartment names and other parameters that have been predefined in parameters
  • constant (constant): parameter is constant, set a fixed number of individuals moving per time step
  • transitionProb (transition probability): parameter is transitionProb, a fix rate transition moving per time step
  • nonparametric: parameter is waitingTime, put a vector contains the waiting time values, the values can be percentages or frequencies, for example {"distribution": "nonparametric", "waitingTime": [3, 5.323, 2, 2.02, 3, 3, 1]}