Application: Logistics Problem

Continuing what you learned from the last chapter, here you'll learn how to help a distributor that's trying to optimize how they combine their products into shipments. You'll put everything you've learned into action, and pick up a bunch of new details along the way.

Task

1. Understand the Goals
   Analyze the program's objectives and refer to the video for guidance and inspiration;

2. Locate the Excel File
   Identify the file containing data about bins and industry grades for palettes;

3. Import the Data
   Use a library like Pandas to load the Excel file, extracting data on bins and industry grading standards;

4. Generate Bin Combinations
   Create all possible combinations of bins for grouping them into palettes, as demonstrated in the last chapter;

5. Evaluate Each Combination
   For every possible bin combination:

   • Analyze Each Palette: compute average properties (e.g., tensile strength and percent defects) for the bins in each palette;
   • Grade and Price Palettes: assign grades and calculate prices based on the averages and industry standards from the Excel file;
   • Compare Sales Prices: track the best combination by comparing the total sales price of the current combination with previous results;

6. Identify the Best Combination
   Once the optimal combination is determined, extract the corresponding bin tags using row indices;

7. Export Results
   Save the final results, including the best combination and associated details, back to an Excel file;

8. Quality Control

   • Cross-check the program's output with the results shown in the video, ensuring the optimal price matches;
   • Validate palette properties against industry standards using spreadsheet calculations;

9. Iterate and Refine

   • Design your program with a modular approach for better organization;
   • Adjust and test your implementation iteratively to ensure accuracy and reliability.

1. Data Importing

   • Bin Properties: import as a matrix containing properties such as weight, tensile strength, and percent defects;
   • Bin Tags: import as a separate matrix;
   • Industry Grade Standards: import as a matrix containing minimal tensile strength, maximum percent defects, and price per palette of 3 bins (7500 lbs);
   • Industry Grade Names: Import as a cell array;

2. Generate Combinations

   • Instead of using Generate_Combinations_MMS_M from chapter 3, use the perms function to generate permutations directly;

3. Identifying Bin Tags

   • Bin tags are recorded as indices that indicate the row positions in the original data. Convert these indices into bin tags using the row indices from the bin tags matrix;
   • Ensure that the row indices are correctly matched between the bin tags and the original data;

4. Handling Dimensions and Indices

   • 2D Matrices: these are used for importing and exporting data to and from Excel. Make sure to reference the correct rows and columns;
   • 3D Matrices: the palette_permutations matrix contains all possible bin combinations folded into a 3D matrix;
     • Each row represents a specific combination of bins into palettes;
     • Each column represents the index of a specific bin;
     • The third dimension (1, 2, 3) corresponds to different palettes;

5. Divide and Conquer Approach

   • Limit the for loop to a single iteration (e.g., for 1:1) to finish the rest of the program and output initial results;
   • Focus on getting the program to output bin tags, palette grades, and the optimal price to Excel one at a time. You can comment out parts of the code to focus on specific aspects;

6. Verification

   • Manually verify the average properties of each palette to ensure they are correctly calculated and graded, as well as the total price of the palette combination;
   • If issues arise, use these verifications to diagnose problems within the for loop;

7. Test Specific Permutations

   • If results are correct for one permutation but an optimal combination isn't found, limit the for loop to test a specific permutation, such as for 32280:32280 or for 16640:16640. This allows you to check performance on drastically different combinations;

8. Troubleshooting

   • If the issue persists after verifying different permutations, there may be a problem with the logic that selects the best permutation from the evaluated iterations. Check the video to compare your results and ensure accuracy.