如何高效生成随机矩阵并执行矩阵乘法？附空列表追加实现

Hey Shahid! Let's fix up your matrix multiplication code using list appends, and also cover smarter ways to generate random matrices and optimize the multiplication itself.

First: Fixing the Pure Python Implementation (List Append Style) #

Your existing code has a few gaps (like an uninitialized res_matrix and missing mat2 setup). Here's a complete, working version that sticks to the list append approach you requested:

import random

# Define the size of your square matrices (adjust this value as needed)
order = 3

# Generate random matrix 1 using list appends
mat1 = []
for _ in range(order):
    # Use randint instead of sample to avoid errors when order > 9 (sample requires unique elements)
    row = [random.randint(1, 9) for _ in range(order)]
    mat1.append(row)
print("Matrix 1:")
for row in mat1:
    print(row)

# Generate random matrix 2 the same way
mat2 = []
for _ in range(order):
    row = [random.randint(1, 9) for _ in range(order)]
    mat2.append(row)
print("\nMatrix 2:")
for row in mat2:
    print(row)

# Initialize result matrix with zeros using list appends
res_matrix = []
for _ in range(len(mat1)):
    # Create a row of zeros matching the number of columns in mat2
    res_row = [0] * len(mat2[0])
    res_matrix.append(res_row)

# Perform matrix multiplication
print("\nResult of Matrix Multiplication:")
for p in range(len(mat1)):
    for q in range(len(mat2[0])):
        for r in range(len(mat2)):
            res_matrix[p][q] += mat1[p][r] * mat2[r][q]

# Print the final result
for res_row in res_matrix:
    print(res_row)

Key Notes on This Implementation: #

• Random Matrix Generation: I swapped random.sample for random.randint because sample requires all elements to be unique—if your order is larger than 9 (the range you used: 1-9), this will throw an error. randint allows duplicate values (standard for random matrices) and works for any order size.
• Result Matrix Initialization: We explicitly build res_matrix by appending rows of zeros, which ensures we have a valid structure to store multiplication results (your original code tried to modify res_matrix before initializing it, which would cause a NameError).

For Better Efficiency: Use NumPy #

If you're working with larger matrices, pure Python loops will get slow quickly. NumPy is built for fast numerical operations and handles matrix multiplication natively. Here's how to do it with NumPy:

import numpy as np

order = 3

# Generate random matrices (values 1-9, integer type)
mat1 = np.random.randint(1, 10, size=(order, order))
mat2 = np.random.randint(1, 10, size=(order, order))

# Matrix multiplication (use the @ operator for clean syntax, or np.dot for older NumPy versions)
res_matrix = mat1 @ mat2

print("Matrix 1:\n", mat1)
print("\nMatrix 2:\n", mat2)
print("\nResult:\n", res_matrix)

This is exponentially faster for large matrices because NumPy offloads the heavy lifting to optimized C code, avoiding Python's slow loop overhead.

内容的提问来源于stack exchange，提问作者Shahid J.