QOPT OPTIMIZATION LIBRARY

Version 0.17.0
Developed by Joseph Hobbs under the MIT License

SYNTAX

    [binary-name] [--flag]* [--parameter value]*

FLAGS

    --opt-help              Displays this help menu
    --quiet                 Does not print updates each iteration
    --no-stop-early         Disables gradient-based convergence criterion

PARAMETERS

    DESCRIPTIONS

        --print-every [int]     Prints updates every `[int]` iterations
        --paradigm [str]        Optimizes using the given paradigm
        --criterion [float]     Sets the magnitude for gradient-based convergence criterion (default: 1.0e-5)
        --maxiter [int]         Sets the maximum number of iterations (default: 100)
        --maxtemp [float]       Sets the maximum annealing temperature (for `simulated-annealing` only)
        --stdev [float]         Sets the standard deviation for population mutation (for stochastic methods only)
        --threads [int]         Sets the number of CPU threads over which to conduct optimization

    DEFAULTS

        --print-every           1
        --paradigm [str]        steepest-descent
        --criterion [float]     1.0e-3
        --maxiter [int]         100
        --maxtemp [float]       1.0
        --stdev [float]         1.0
        --threads               1

PARADIGMS

    DETERMINISTIC METHODS

        `steepest-descent` (default)
            Steepest descent using a gradient-based method.  Ensure to set `criterion` appropriately.
        
        `newton`
            Newton's method.  Ensure to set `criterion` appropriately.

    STOCHASTIC METHODS

        `genetic`
            Genetic algorithm.  Ensure to set `maxiter` and `stdev` appropriately.

        `simulated-annealing`
            Simulated annealing.  Ensure to set `maxiter`, `maxtemp`, and `stdev` appropriately.
