Project Overview

This is a rebar cutting stock optimization system that solves the cutting stock problem for steel reinforcement bars using depth-first search algorithms and integer linear programming.

Commands

Running the Application

# Run the Streamlit web interface
streamlit run cutting_optimizer_streamlit.py

# Run command-line optimizer
python cutting_optimizer_ver2.py

Dependencies

The system requires these Python packages:

• streamlit: Web interface
• pulp: Integer linear programming solver
• openpyxl: Excel file processing
• pandas: Data processing

Testing with Different Diameters

Modify the diameter variable in the main functions (typically 'D10', 'D13', 'D16', 'D19', 'D22').

Architecture

Core Components

1. Optimizers:

   • cutting_optimizer_streamlit.py: Main web interface with integrated optimization
   • cutting_optimizer_ver2.py: Command-line version with Excel integration

2. Data Processing:

   • read_xlsx.py: Excel parser for cutting requirements with multi-sheet support
   • result_sheet.py: Results output to Excel format

Key Algorithms

• Depth-First Search (DFS): Generates all valid cutting combinations within material constraints using recursive exploration
• Integer Linear Programming (ILP): Optimizes cutting patterns using PuLP library to minimize waste

Rebar Specifications

The system supports five standard rebar diameters with predefined stock lengths:

BASE_PATTERNS = {
    'D10': [4000, 4500, 5500, 6000],
    'D13': [3000, 4000, 4500, 5500, 6000, 7500],
    'D16': [4000, 4500, 5500, 6000, 7000],
    'D19': [3500, 4000, 4500, 5500, 6000],
    'D22': [4000, 4500, 5500, 6000]
}

Data Flow

1. Input: Excel files with cutting requirements (required_cuts.xlsx)
2. Processing: Extract rebar diameter data using find_cell_position() and get_diameter_column_index()
3. Combination Generation: Use DFS to find valid cutting patterns within stock constraints
4. Optimization: Solve ILP problem with 120-second timeout to minimize waste
5. Output: Display cutting instructions with yield rates and material usage statistics

File Structure

• Input files: Excel files with "鉄筋径" (rebar diameter) headers
• Result files: Located in result/ directory with processing results and comparisons

Performance Considerations

• Combination generation is computationally expensive for large problems
• Default optimization timeout: 120 seconds
• Results include processing time and yield rate calculations
• DFS uses pruning to avoid exploring invalid branches when current_sum > max_sum