random_numbers.md

marp	math	theme	footer
true	katex	custom-theme

Random number generation

Today:

• How to generate random numbers in modern C++
• What's wrong with rand()

📺 Watch the related YouTube video!

---

Special symbols used in slides

• 🎨 - Style recommendation
• 🎓 - Software design recommendation
• 😱 - Not a good practice! Avoid in real life!
• ✅ - Good practice!
• ❌ - Whatever is marked with this is wrong
• 🚨 - Alert! Important information!
• 💡 - Hint or a useful exercise
• 🔼1️⃣7️⃣ - Holds for this version of C++(here, 17) and above
• 🔽1️⃣1️⃣ - Holds for versions until this one C++(here, 11)

Style (🎨) and software design (🎓) recommendations mostly come from Google Style Sheet and the CppCoreGuidelines

---

What are random numbers?

---

What are random numbers?

• Random numbers should be hard to predict (under the definition of "hard")
• Mostly, we generate pseudo-random numbers instead
• There are multiple algorithms for this:
  • Linear congruential generator - Fast and small storage
  • Lagged Fibonacci generator - Very fast, uses more storage
  • Mersenne Twister - Slower, uses more storage but generates very high quality pseudo random numbers
• We will not go into details here!
• If you want to learn more, you can:
  • Read a book Random Numbers and Computers
  • Read this technical report for a slightly advanced intro

---

But how do I use it?

• 🔼1️⃣1️⃣ Include <random>
• ✅ There is a great summary of pseudo random number generation on cppreference.com
• Implements all the algorithms listed on previous slide
• Here is a summary of how it works:
  1. We need a "random device" that is our source of non-deterministic uniform (maybe pseudo-) random numbers
  2. We pass this device into a "random number engine" from the previous slide
  3. We pass this engine into a "random number distribution" which generates the resulting numbers we want

---

Let's see it in practice

#include <iostream>
#include <random>

int main() {
  std::random_device random_device;
  std::mt19937 random_engine{random_device()};
  std::uniform_real_distribution distribution{23.0, 42.0};
  for (auto i = 0; i < 5; ++i) {
    std::cout << distribution(random_engine) << std::endl;
  }
  return 0;
}

This gives an output that looks smth like this:

33.9856
30.8976
40.8357
37.9964
27.8459

---

What's wrong with rand()?

• It is available from the <cstdlib> header
• rand() is a C function (not C++)
• The only option before C++11
• 😱 Let's see how it's typically used:

  // Somewhere in main
  std::srand(seed);  // Initialize random seed
  int random_variable = std::rand();

• 😱 It uses global state (seed is set globally)
• The quality of the generated random sequence is not guaranteed
• ✅ Always use methods from <random> instead!

---