FedEx ULD Packer

Installation

1. Install Python and C++

2. Install Gurobi

   • Download Gurobi Optimizer from Gurobi's official website
   • Install the Gurobi Optimizer following their platform-specific instructions
   • Obtain a Gurobi license:
     • Academic users can get a free license
     • Commercial users need to purchase a license
     • We have the licence but due to inter IIT rules we cant share the key, if any assistance is needed you are free to contact us

   Activate the Gurobi license:

   grbgetkey YOUR_LICENSE_KEY  # Replace with your actual license key

3. Set up Virtual Environment and Install Requirements

   python3 -m venv packer_env
   source packer_env/bin/activate  # On Windows, use `venv\Scripts\activate`
   pip install -r requirements.txt

Usage

Running the Packer

1. Add input data in challenge_input.txt

2. Run

python fit.py

Default runtime is 35 min, can be modified by passing --timelimit=time_in_minutes while running fit.py or modifying fit.py itself

3. Get output in challenge_output.txt

Running the Visualizer

1. Run

python visualizer.py

It plots challenge_output.txt and saves the images in Visuals directory

Directory Breakdown

Main Runner

• fit.py:
  • Primary script to execute the entire optimization process
  • Coordinates different components of the packing algorithm
  • Handles high-level workflow and integration of various modules
  • Time limit for running the algorithm, can be configured here ( default is 35 mins ).
  • Run commands for python/C++ can be also be configured here ( default is python and g++ ).

Python Scripts

• fillPriority.py:
  • Focuses on priority-based packing algorithm
  • Ensures high-priority items are packed with specific constraints
• fillEconomyWithPriority.py:
  • Advanced filling strategy combining economy considerations with priority constraints
  • Fill using Volume cuts and API
• fillEconomy.py:
  • Implements economy-class packing strategy
  • Fill using metric cuts and API
• runMIP.py:
  • Executes Linear Programming Problem (MIP)
  • Manages optimization solver configuration and problem setup
• makeDiscrete.py:
  • Converts continuous variables to discrete values
  • Helps in transforming optimization results to practical, implementable packing solutions

C++ Conversion Utilities

• input_converter.cpp:
  • Converts input data into a format suitable for optimization
  • Preprocesses and transforms input files
• output_converter.cpp:
  • Converts optimization results into desired output format
  • Processes and transforms solver output
• output_checker.cpp:
  • Validates the output of the optimization process
  • Performs checks to ensure solution feasibility and quality
• gravity.cpp:
  • Implements gravity-based algorithms and settling

I/O Files

• challenge_input.txt:
  • Contains initial input data for the optimization challenge
  • Provides problem specifications and constraints
• challenge_output.txt:
  • Contains final output data for the optimization challenge

Utilities

• visualizer.py:
  • Visualize the output of packaging
  • Output is stored in /Visuals
• apiCred.txt
  • Contains api username and apikey from 3dbinpacking.com

Visuals/

• Directory used for storing final visuals

Logs/

• Directory used for storing intermediate files generated during pipeline execution

Generators/

• APIinitialise.py:
  • Initializes API connections and configurations
  • Sets up communication with external data sources or services
• APIresponseConvert.py:
  • Processes and converts API responses
  • Transforms raw API data into usable format for optimization
• discrete.py:
  • Implements discretization
• generateConstants.py:
  • Generates constant values used in optimization
  • Provides fixed parameters for the packing algorithm
• generateInitials.py:
  • Helps in setting up the initial state of the packing problem
• generateVariables.py:
  • Generates variables for the optimization model
  • Defines decision variables used in the packing algorithm
• GurobiBatchVariables.py:
  • Manages batch processing of Gurobi variables
  • Handles large-scale variable generation and management
• GurobiConstraints.py:
  • Defines constraints for Gurobi optimization model
  • Implements packing rules and restrictions
• GurobiInitialise.py:
  • Initializes Gurobi optimization environment
  • Sets up the Gurobi solver with necessary configurations
• GurobiStart.py:
  • Creates starting point for Gurobi Variables
• printState.py:
  • Utility for printing current state of optimization
  • Helps in debugging and monitoring the optimization process
• readConstants.py:
  • Reads predefined constants from configuration files
  • Loads fixed parameters for the optimization
• readInitials.py:
  • Reads initial conditions from external sources
  • Loads starting points for the optimization
• readVariables.py:
  • Reads variables from external sources
  • Loads variable configurations for the optimization model

tried_approaches/

Codes for the tried approaches, MIP with Batching and Metric, and both Greedy Approaches - Randomized Greedy Search with Metric and Maximal Surface Area with Metric. Each Directory has its own README.md for its usage.

• batch_mip/:
  • Code for MIP with Batching and Metric.
• maximal_surface_area/
  • Code for Greedy Approach 1 - Maximal Surface Area with Metric
• randomized_greedy_approach/
  • Code for Greedy Approach 2 - Randomized Greedy Search with Metric