R语言有向图随机游走模拟函数实现求助

Hey there! Since you're new to R and working with Markov chains (which this directed graph + transition matrix setup is), let's break this down into simple, actionable steps. We'll build a function that simulates a random walk through your nodes and outputs the visit probabilities you need.

Step 1: Define Your Transition Matrix #

First, let's translate your given matrix into R code. We'll use the matrix() function and add clear labels for readability:

# Define the 6x6 transition probability matrix
transition_matrix <- matrix(
  c(0.0, 0.5, 0.3, 0.0, 0.0, 0.2,
    0.1, 0.2, 0.0, 0.4, 0.1, 0.2,
    0.5, 0.0, 0.0, 0.0, 0.0, 0.5,
    0.0, 0.1, 0.0, 0.0, 0.6, 0.3,
    0.0, 0.0, 0.0, 0.4, 0.0, 0.6,
    0.4, 0.0, 0.0, 0.0, 0.2, 0.4),
  nrow = 6,
  byrow = TRUE,
  dimnames = list(paste0("Node ", 1:6), paste0("Node ", 1:6))
)

# Check the matrix to make sure it's correct
print(transition_matrix)

Step 2: Build the Simulation Function #

This function will simulate a random walk through the nodes for your specified number of steps, count visits to each node, and calculate the probabilities. I'll add comments to explain each part so you can follow along:

calculate_visit_probs <- function(transition_matrix, steps) {
  # Get the number of nodes from the matrix dimensions
  num_nodes <- nrow(transition_matrix)
  
  # Initialize a counter to track visits to each node (starts at 0 for all)
  visit_counts <- rep(0, num_nodes)
  
  # Start at a random node (you can also fix this to Node 1 if you prefer)
  current_node <- sample(1:num_nodes, 1)
  
  # Simulate the random walk for the given number of steps
  for (step in 1:steps) {
    # Increment the visit count for the current node
    visit_counts[current_node] <- visit_counts[current_node] + 1
    
    # Choose the next node using the transition probabilities from current node
    current_node <- sample(1:num_nodes, 1, prob = transition_matrix[current_node, ])
  }
  
  # Convert counts to probabilities by dividing by total steps
  visit_probs <- visit_counts / steps
  
  # Return the probability vector
  return(visit_probs)
}

Quick Explanation of the Function: #

• sample(1:num_nodes, 1, prob = ...): This picks the next node randomly, weighted by the transition probabilities from the current node—exactly how a Markov chain works!
• We count each visit to a node, then divide by total steps to get the proportion of time spent at each node (your visit probability).

Step 3: Use the Function #

Now let's test it with your requested 100,000 steps:

# Run the simulation with 100,000 steps
node_probs <- calculate_visit_probs(transition_matrix, steps = 100000)

# Print the result (labels make it easier to read)
names(node_probs) <- paste0("Node ", 1:6)
print(node_probs)

Example Output: #

You'll get something close to your sample output (since it's a simulation, results vary slightly each run):

Node 1      Node 2      Node 3      Node 4      Node 5      Node 6 
0.18492000 0.13057000 0.05523000 0.11381000 0.15379000 0.36168000

Notes for New R Users #

• Simulation vs. Exact Steady State: This simulation gives you an approximation of the steady-state distribution (the long-run visit probabilities). If you want the exact steady state, you could solve a system of linear equations, but simulation is more intuitive for beginners.
• Initial Node: Starting at a random node ensures we don't bias the results. If you start at Node 1 every time, the first few steps will be skewed, but with 100,000 steps, the initial node's impact becomes negligible.

内容的提问来源于stack exchange，提问作者Md Moshraf Hossain