R语言模拟100条一维粒子随机运动路径报错及可视化求助

Got it, let's work through your code issues and get those particle paths visualized properly. First, let's fix that "replacement has zero length" error, then cover how to plot the paths clearly.

Why You're Getting the Error #

Your code has a few key issues that are causing the error and incorrect behavior:

• You didn't initialize the x array before trying to assign values to x[j] or x[0]—R can't assign to a non-existent object.
• R uses 1-based indexing (not 0-based), so x[0] = 0 is invalid and will cause problems.
• Your W generation logic is off: for Brownian motion with drift, the random increment should have a mean of 0 (the drift is handled separately with Dt*V), and you need one random value per particle per time step.

Fixed Code to Simulate Particle Paths #

Let's rewrite the code to correctly track 100 particles over 100 time steps:

# Define simulation parameters
n_particles <- 100  # Number of particles
n_steps <- 100      # Number of time steps
Dt <- 0.0001        # Time step (0.1ms = 0.0001 seconds)
V <- 0.5            # Drift velocity

# Initialize a matrix to store paths: rows = particles, columns = time steps
# The first column is the initial position (0 for all particles)
x_paths <- matrix(0, nrow = n_particles, ncol = n_steps + 1)

# Simulate each time step
for (step in 1:n_steps) {
  # Generate random increments for all particles: mean=0, sd=sqrt(Dt) (Brownian motion)
  W <- rnorm(n_particles, mean = 0, sd = sqrt(Dt))
  # Update position: previous position + drift term + random term
  x_paths[, step + 1] <- x_paths[, step] + Dt * V + W
}

This code:

• Uses a matrix to neatly store all particle positions (easy to access later)
• Avoids 0-indexing issues by starting at step 1
• Generates the correct random increments for each particle at every time step

Visualizing the Paths #

You can use either base R plotting or ggplot2 for more polished visuals. Here are both options:

Option 1: Base R Plot #

# Create a time vector matching our steps
time <- seq(0, n_steps * Dt, by = Dt)

# Plot all particle paths (semi-transparent to avoid clutter)
plot(time, x_paths[1, ], type = "l", col = rgb(0, 0, 1, 0.2),
     xlab = "Time (seconds)", ylab = "Position",
     main = "1D Random Walk with Drift (100 Particles)")

# Add the rest of the particle paths
for (i in 2:n_particles) {
  lines(time, x_paths[i, ], col = rgb(0, 0, 1, 0.2))
}

# Add a red line for the mean position of all particles
mean_path <- colMeans(x_paths)
lines(time, mean_path, col = "red", lwd = 2)

# Add a legend
legend("topleft", legend = c("Individual Particle", "Mean Path"),
       col = c(rgb(0, 0, 1, 0.2), "red"), lwd = c(1, 2))

Option 2: ggplot2 (Cleaner, More Customizable) #

First, we'll reshape the data into a long format (required for ggplot2):

library(ggplot2)
library(tidyr)

# Convert the path matrix to a tidy data frame
x_df <- as.data.frame(x_paths) %>%
  mutate(particle = 1:n_particles) %>%
  pivot_longer(cols = -particle, names_to = "step", values_to = "position") %>%
  mutate(time = as.numeric(substr(step, 2, nchar(step))) * Dt)

# Plot with ggplot2
ggplot(x_df, aes(x = time, y = position, group = particle)) +
  geom_line(color = "steelblue", alpha = 0.2) +  # Semi-transparent particle paths
  geom_line(aes(y = mean(position), group = NULL), 
            color = "firebrick", size = 1) +  # Mean path
  labs(x = "Time (seconds)", y = "Position",
       title = "1D Random Walk with Drift (100 Particles)") +
  theme_minimal()

The semi-transparent blue lines let you see the overall spread of particles, while the red mean line highlights the drift effect from your V parameter.

内容的提问来源于stack exchange，提问作者Pablodour