Skip to content

README_USAGE.md

Latest commit

 

History

History
108 lines (78 loc) · 3.46 KB

README_USAGE.md

File metadata and controls

108 lines (78 loc) · 3.46 KB
Raw

ECHO Offline Optimizer - Usage Guide

Command-Line Usage

The compiled echo_offline_debug executable accepts 6 command-line arguments:

./echo_offline_debug [nPop] [generations] [mutationRate] [elite] [inertia] [nCycles]

Parameters

Parameter Type Range Default Description
nPop integer 10-10000 100 Population size for genetic algorithm
generations integer 1-1000 10 Number of generations to evolve
mutationRate float 0.0-1.0 0.1 Probability of mutation per gene
elite float 0.0-1.0 0.3 Proportion of elite individuals preserved
inertia float 0.0-1.0 0.3 Proportion of population replaced each cycle
nCycles integer 1-100 5 Number of PCS optimization cycles

Examples

# Default parameters
./echo_offline_debug

# Small quick run (development)
./echo_offline_debug 50 5 0.1 0.3 0.3 2

# Larger optimization (production)
./echo_offline_debug 200 50 0.15 0.25 0.4 10

# Get help
./echo_offline_debug --help

Jupyter Notebook Usage

The plot_echo_results.ipynb notebook has been configured to:

  1. Automatically compile the C code on the first cell
  2. Run the optimizer with customizable parameters in the second cell
  3. Plot and analyze results in subsequent cells

Workflow

  1. Open plot_echo_results.ipynb in Jupyter
  2. Run the first cell to compile echo_offline_debug.c
  3. Modify the parameter values in the second cell: 
    nPop = 100            # Adjust population size
generations = 10      # Adjust iterations
mutation_rate = 0.1   # Adjust mutation rate
elite_prop = 0.3      # Adjust elite preservation
inertia = 0.3         # Adjust population turnover
n_cycles = 5          # Adjust number of cycles
  4. Run the second cell to execute the optimizer
  5. Run remaining cells to generate plots and analysis

Output Files

After execution, the following CSV files are generated:

  • profile_output.csv - Final optimized profile vs target

    • Columns: rho, target, achieved, gyro1, gyro2, ... gyroN
    • Shows power distribution from each gyrotron

  • profile_output_all_cycles.csv - All intermediate results

    • Columns: cycle, rho, target, achieved, gyro1, gyro2, ... gyroN
    • Allows tracking optimization progress

  • gyro_parameters.csv - Final angles and duty cycles

    • Columns: Gyrotron, Angle, DutyCycle, PowerSupply
    • Ready for experimental implementation

Compilation

To manually recompile:

gcc -o echo_offline_debug echo_offline_debug.c \
    echo/echo_torbeamNN_x.c echo/echo_torbeamNN_o.c \
    k2c/*.c -lm

Performance Tips

  • Quick testing: Use small nPop (50) and generations (5)
  • Publication-ready: Use large nPop (200+) and generations (50+)
  • Multi-parameter study: Run notebook multiple times with different parameters
  • Parallel runs: Execute multiple notebook instances with different configs

Troubleshooting

"Command not found: ./echo_offline_debug"

  • Ensure you're in the /Users/ar0535/Downloads/echo_offline directory
  • Recompile with: gcc -o echo_offline_debug echo_offline_debug.c echo/echo_torbeamNN_x.c echo/echo_torbeamNN_o.c k2c/*.c -lm

CSV files not created

  • Check file permissions in current directory
  • Ensure profile_output.csv in root directory (not subdirectory)

Segmentation fault

  • Memory issue likely from very large population sizes
  • Reduce nPop and rerun