boost/math
2
0
Fork 0
mirror of https://github.com/boostorg/math.git synced 2026-01-19 04:22:09 +00:00
Code Issues Projects Releases Wiki Activity
Files
develop
math/doc/html/math_toolkit/differential_evolution.html
Tomato-in af8ba3f7b2 Adjust parameter order in Differential Evolution algorithm 
Problem Description
The optimization module implements four algorithms: Differential
Evolution, Algorithm jSO, Random Search, and Evolution Strategy with
Covariance Matrix Adaptation. All these algorithms accept seven
parameters: `cost_function`, `params`, `gen`, `value_to_reach`,
`cancellation`, `current_minimum_cost`, and `queries`. Unfortunately,
the Differential Evolution algorithm doesn't maintain parameter
consistency with the other three algorithms: the last two parameters
(`current_minimum_cost` and `queries`) are in reverse order compared to
the other algorithms. This appears to be an unintentional oversight.

Solution
- Adjust the parameter order in the Differential Evolution algorithm to ensure consistency with other optimization algorithms
- Update the algorithm documentation accordingly
(No changes to example code were needed as the examples don't use these two parameters)
2025-05-15 07:19:58 -07:00

267 lines
26 KiB
HTML
Raw Permalink Blame History

This file contains invisible Unicode characters
This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<title>Differential Evolution</title>
<link rel="stylesheet" href="../math.css" type="text/css">
<meta name="generator" content="DocBook XSL Stylesheets V1.79.1">
<link rel="home" href="../index.html" title="Math Toolkit 4.2.1">
<link rel="up" href="../optimization.html" title="Chapter 11. Optimization">
<link rel="prev" href="../optimization.html" title="Chapter 11. Optimization">
<link rel="next" href="jso.html" title="Algorithm jSO">
<meta name="viewport" content="width=device-width, initial-scale=1">
</head>
<body bgcolor="white" text="black" link="#0000FF" vlink="#840084" alink="#0000FF">
<table cellpadding="2" width="100%"><tr>
<td valign="top"><img alt="Boost C++ Libraries" width="277" height="86" src="../../../../../boost.png"></td>
<td align="center"><a href="../../../../../index.html">Home</a></td>
<td align="center"><a href="../../../../../libs/libraries.htm">Libraries</a></td>
<td align="center"><a href="http://www.boost.org/users/people.html">People</a></td>
<td align="center"><a href="http://www.boost.org/users/faq.html">FAQ</a></td>
<td align="center"><a href="../../../../../more/index.htm">More</a></td>
</tr></table>
<hr>
<div class="spirit-nav">
<a accesskey="p" href="../optimization.html"><img src="../../../../../doc/src/images/prev.png" alt="Prev"></a><a accesskey="u" href="../optimization.html"><img src="../../../../../doc/src/images/up.png" alt="Up"></a><a accesskey="h" href="../index.html"><img src="../../../../../doc/src/images/home.png" alt="Home"></a><a accesskey="n" href="jso.html"><img src="../../../../../doc/src/images/next.png" alt="Next"></a>
</div>
<div class="section">
<div class="titlepage"><div><div><h2 class="title" style="clear: both">
<a name="math_toolkit.differential_evolution"></a><a class="link" href="differential_evolution.html" title="Differential Evolution">Differential Evolution</a>
</h2></div></div></div>
<h4>
<a name="math_toolkit.differential_evolution.h0"></a>
<span class="phrase"><a name="math_toolkit.differential_evolution.synopsis"></a></span><a class="link" href="differential_evolution.html#math_toolkit.differential_evolution.synopsis">Synopsis</a>
</h4>
<pre class="programlisting"><span class="preprocessor">#include</span> <span class="special">&lt;</span><span class="identifier">boost</span><span class="special">/</span><span class="identifier">math</span><span class="special">/</span><span class="identifier">optimization</span><span class="special">/</span><span class="identifier">differential_evolution</span><span class="special">.</span><span class="identifier">hpp</span><span class="special">&gt;</span>
<span class="keyword">namespace</span> <span class="identifier">boost</span><span class="special">::</span><span class="identifier">math</span><span class="special">::</span><span class="identifier">optimization</span> <span class="special">{</span>
<span class="keyword">template</span> <span class="special">&lt;</span><span class="keyword">typename</span> <span class="identifier">ArgumentContainer</span><span class="special">&gt;</span> <span class="keyword">struct</span> <span class="identifier">differential_evolution_parameters</span> <span class="special">{</span>
<span class="keyword">using</span> <span class="identifier">Real</span> <span class="special">=</span> <span class="keyword">typename</span> <span class="identifier">ArgumentContainer</span><span class="special">::</span><span class="identifier">value_type</span><span class="special">;</span>
<span class="identifier">ArgumentContainer</span> <span class="identifier">lower_bounds</span><span class="special">;</span>
<span class="identifier">ArgumentContainer</span> <span class="identifier">upper_bounds</span><span class="special">;</span>
<span class="identifier">Real</span> <span class="identifier">mutation_factor</span> <span class="special">=</span> <span class="keyword">static_cast</span><span class="special">&lt;</span><span class="identifier">Real</span><span class="special">&gt;(</span><span class="number">0.65</span><span class="special">);</span>
<span class="keyword">double</span> <span class="identifier">crossover_probability</span> <span class="special">=</span> <span class="number">0.5</span><span class="special">;</span>
<span class="comment">// Population in each generation:</span>
<span class="identifier">size_t</span> <span class="identifier">NP</span> <span class="special">=</span> <span class="number">200</span><span class="special">;</span>
<span class="identifier">size_t</span> <span class="identifier">max_generations</span> <span class="special">=</span> <span class="number">1000</span><span class="special">;</span>
<span class="identifier">size_t</span> <span class="identifier">threads</span> <span class="special">=</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">thread</span><span class="special">::</span><span class="identifier">hardware_concurrency</span><span class="special">();</span>
<span class="identifier">ArgumentContainer</span> <span class="keyword">const</span> <span class="special">*</span> <span class="identifier">initial_guess</span> <span class="special">=</span> <span class="keyword">nullptr</span><span class="special">;</span>
<span class="special">};</span>
<span class="keyword">template</span> <span class="special">&lt;</span><span class="keyword">typename</span> <span class="identifier">ArgumentContainer</span><span class="special">,</span> <span class="keyword">class</span> <span class="identifier">Func</span><span class="special">,</span> <span class="keyword">class</span> <span class="identifier">URBG</span><span class="special">&gt;</span>
<span class="identifier">ArgumentContainer</span> <span class="identifier">differential_evolution</span><span class="special">(</span>
<span class="keyword">const</span> <span class="identifier">Func</span> <span class="identifier">cost_function</span><span class="special">,</span> <span class="identifier">differential_evolution_parameters</span><span class="special">&lt;</span><span class="identifier">ArgumentContainer</span><span class="special">&gt;</span> <span class="keyword">const</span> <span class="special">&amp;</span><span class="identifier">de_params</span><span class="special">,</span> <span class="identifier">URBG</span> <span class="special">&amp;</span><span class="identifier">g</span><span class="special">,</span>
<span class="identifier">std</span><span class="special">::</span><span class="identifier">invoke_result_t</span><span class="special">&lt;</span><span class="identifier">Func</span><span class="special">,</span> <span class="identifier">ArgumentContainer</span><span class="special">&gt;</span> <span class="identifier">target_value</span> <span class="special">=</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">numeric_limits</span><span class="special">&lt;</span><span class="identifier">std</span><span class="special">::</span><span class="identifier">invoke_result_t</span><span class="special">&lt;</span><span class="identifier">Func</span><span class="special">,</span> <span class="identifier">ArgumentContainer</span><span class="special">&gt;&gt;::</span><span class="identifier">quiet_NaN</span><span class="special">(),</span>
<span class="identifier">std</span><span class="special">::</span><span class="identifier">atomic</span><span class="special">&lt;</span><span class="keyword">bool</span><span class="special">&gt;</span> <span class="special">*</span><span class="identifier">cancellation</span> <span class="special">=</span> <span class="keyword">nullptr</span><span class="special">,</span>
<span class="identifier">std</span><span class="special">::</span><span class="identifier">atomic</span><span class="special">&lt;</span><span class="identifier">std</span><span class="special">::</span><span class="identifier">invoke_result_t</span><span class="special">&lt;</span><span class="identifier">Func</span><span class="special">,</span> <span class="identifier">ArgumentContainer</span><span class="special">&gt;&gt;</span> <span class="special">*</span><span class="identifier">current_minimum_cost</span> <span class="special">=</span> <span class="keyword">nullptr</span><span class="special">,</span>
<span class="identifier">std</span><span class="special">::</span><span class="identifier">vector</span><span class="special">&lt;</span><span class="identifier">std</span><span class="special">::</span><span class="identifier">pair</span><span class="special">&lt;</span><span class="identifier">ArgumentContainer</span><span class="special">,</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">invoke_result_t</span><span class="special">&lt;</span><span class="identifier">Func</span><span class="special">,</span> <span class="identifier">ArgumentContainer</span><span class="special">&gt;&gt;&gt;</span> <span class="special">*</span><span class="identifier">queries</span> <span class="special">=</span> <span class="keyword">nullptr</span><span class="special">);</span>
<span class="special">}</span> <span class="comment">// namespaces</span>
</pre>
<p>
The <code class="computeroutput"><span class="identifier">differential_evolution</span></code>
function provides an implementation of the (classical) differential evolution
optimization algorithm, often going by the label <code class="computeroutput"><span class="identifier">de</span><span class="special">/</span><span class="identifier">rand</span><span class="special">/</span><span class="identifier">bin</span><span class="special">/</span><span class="number">1</span></code>.
It is capable of minimizing a cost function defined on a continuous space represented
by a set of bounds. This function has been designed more for progress observability,
graceful cancellation, and post-hoc data analysis than for speed of convergence.
We justify this design choice by reference to the "No free lunch"
theorem of Wolpert and Macready, which establishes "that for any algorithm,
any elevated performance over one class of problems is offset by performance
over another class".
</p>
<h4>
<a name="math_toolkit.differential_evolution.h1"></a>
<span class="phrase"><a name="math_toolkit.differential_evolution.parameters"></a></span><a class="link" href="differential_evolution.html#math_toolkit.differential_evolution.parameters">Parameters</a>
</h4>
<div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; ">
<li class="listitem">
<code class="computeroutput"><span class="identifier">lower_bounds</span></code>: A container
representing the lower bounds of the optimization space along each dimension.
The <code class="computeroutput"><span class="special">.</span><span class="identifier">size</span><span class="special">()</span></code> of the bounds should return the dimension
of the problem.
</li>
<li class="listitem">
<code class="computeroutput"><span class="identifier">upper_bounds</span></code>: A container
representing the upper bounds of the optimization space along each dimension.
It should have the same size of <code class="computeroutput"><span class="identifier">lower_bounds</span></code>,
and each element should be &gt;= the corresponding element of <code class="computeroutput"><span class="identifier">lower_bounds</span></code>.
</li>
<li class="listitem">
<code class="computeroutput"><span class="identifier">mutation_factor</span></code>: Also known
as <code class="computeroutput"><span class="identifier">F</span></code> or scale factor in
the literature. It controls the rate at which the population evolves (default
is 0.65).
</li>
<li class="listitem">
<code class="computeroutput"><span class="identifier">crossover_probability</span></code>:
The crossover probability determining the trade-off between exploration
and exploitation (default is 0.5).
</li>
<li class="listitem">
<code class="computeroutput"><span class="identifier">NP</span></code>: The population size
(default is 200). Parallelization occurs over the population, so this should
be "large".
</li>
<li class="listitem">
<code class="computeroutput"><span class="identifier">max_generations</span></code>: The maximum
number of generations for the optimization (default is 1000).
</li>
<li class="listitem">
<code class="computeroutput"><span class="identifier">threads</span></code>: The number of
threads to use for parallelization (default is the hardware concurrency).
If the objective function is already multithreaded, then this should be
set to 1 to prevent oversubscription.
</li>
<li class="listitem">
<code class="computeroutput"><span class="identifier">initial_guess</span></code>: An optional
guess for where we should start looking for solutions.
</li>
</ul></div>
<p>
The defaults were chosen by a reading of Price, Storn, and Lampinen, chapter
3, with the exception of the population size, which we have chosen a bit larger
than they found as core counts have increased since publication of this reference
and parallelization occurs within each generation. Note that there is a tradeoff
between finding global minima and convergence speed. The most robust way of
decreasing the probability of getting stuck in a local minima is to increase
the population size.
</p>
<h4>
<a name="math_toolkit.differential_evolution.h2"></a>
<span class="phrase"><a name="math_toolkit.differential_evolution.the_function"></a></span><a class="link" href="differential_evolution.html#math_toolkit.differential_evolution.the_function">The
function</a>
</h4>
<pre class="programlisting"><span class="keyword">template</span> <span class="special">&lt;</span><span class="keyword">typename</span> <span class="identifier">ArgumentContainer</span><span class="special">,</span> <span class="keyword">class</span> <span class="identifier">Func</span><span class="special">,</span> <span class="keyword">class</span> <span class="identifier">URBG</span><span class="special">&gt;</span>
<span class="identifier">ArgumentContainer</span> <span class="identifier">differential_evolution</span><span class="special">(</span><span class="keyword">const</span> <span class="identifier">Func</span> <span class="identifier">cost_function</span><span class="special">,</span>
<span class="identifier">differential_evolution_parameters</span><span class="special">&lt;</span><span class="identifier">ArgumentContainer</span><span class="special">&gt;</span> <span class="keyword">const</span> <span class="special">&amp;</span><span class="identifier">de_params</span><span class="special">,</span>
<span class="identifier">URBG</span> <span class="special">&amp;</span><span class="identifier">gen</span><span class="special">,</span>
<span class="identifier">std</span><span class="special">::</span><span class="identifier">invoke_result_t</span><span class="special">&lt;</span><span class="identifier">Func</span><span class="special">,</span> <span class="identifier">ArgumentContainer</span><span class="special">&gt;</span> <span class="identifier">value_to_reach</span>
<span class="special">=</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">numeric_limits</span><span class="special">&lt;</span><span class="identifier">std</span><span class="special">::</span><span class="identifier">invoke_result_t</span><span class="special">&lt;</span><span class="identifier">Func</span><span class="special">,</span> <span class="identifier">ArgumentContainer</span><span class="special">&gt;&gt;::</span><span class="identifier">quiet_NaN</span><span class="special">(),</span>
<span class="identifier">std</span><span class="special">::</span><span class="identifier">atomic</span><span class="special">&lt;</span><span class="keyword">bool</span><span class="special">&gt;</span> <span class="special">*</span><span class="identifier">cancellation</span> <span class="special">=</span> <span class="keyword">nullptr</span><span class="special">,</span>
<span class="identifier">std</span><span class="special">::</span><span class="identifier">atomic</span><span class="special">&lt;</span><span class="identifier">std</span><span class="special">::</span><span class="identifier">invoke_result_t</span><span class="special">&lt;</span><span class="identifier">Func</span><span class="special">,</span> <span class="identifier">ArgumentContainer</span><span class="special">&gt;&gt;</span> <span class="special">*</span><span class="identifier">current_minimum_cost</span> <span class="special">=</span> <span class="keyword">nullptr</span><span class="special">,</span>
<span class="identifier">std</span><span class="special">::</span><span class="identifier">vector</span><span class="special">&lt;</span><span class="identifier">std</span><span class="special">::</span><span class="identifier">pair</span><span class="special">&lt;</span><span class="identifier">ArgumentContainer</span><span class="special">,</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">invoke_result_t</span><span class="special">&lt;</span><span class="identifier">Func</span><span class="special">,</span> <span class="identifier">ArgumentContainer</span><span class="special">&gt;&gt;&gt;</span> <span class="special">*</span><span class="identifier">queries</span> <span class="special">=</span> <span class="keyword">nullptr</span><span class="special">)</span>
</pre>
<p>
Parameters:
</p>
<div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; ">
<li class="listitem">
<code class="computeroutput"><span class="identifier">cost_function</span></code>: The cost
function to be minimized.
</li>
<li class="listitem">
<code class="computeroutput"><span class="identifier">de_params</span></code>: The parameters
to the algorithm as described above.
</li>
<li class="listitem">
<code class="computeroutput"><span class="identifier">rng</span></code>: A uniform random bit
generator, like <code class="computeroutput"><span class="identifier">std</span><span class="special">::</span><span class="identifier">mt19937_64</span></code>.
</li>
<li class="listitem">
<code class="computeroutput"><span class="identifier">value_to_reach</span></code>: An optional
value that, if reached, stops the optimization. This is the most robust
way to terminate the calculation, but in most cases the optimal value of
the cost function is unknown. If it is, use it! See the referenced book
for clear examples of when target values can be deduced.
</li>
<li class="listitem">
<code class="computeroutput"><span class="identifier">cancellation</span></code>: An optional
atomic boolean to allow the user to stop the computation and gracefully
return the best result found up to that point. N.B.: Cancellation is not
immediate; the in-progress generation finishes.
</li>
<li class="listitem">
<code class="computeroutput"><span class="identifier">current_minimum_cost</span></code>: An
optional atomic variable to store the current minimum cost during optimization.
This allows developers to (e.g.) plot the progress of the minimization
over time and in conjunction with the cancellation argument allow halting
the computation when the progress stagnates. Refer to Price, Storn, and
Lampinen, Section 3.2 for caveats with this approach.
</li>
<li class="listitem">
<code class="computeroutput"><span class="identifier">queries</span></code>: An optional vector
to store intermediate results during optimization. This is useful for debugging
and perhaps volume rendering of the objective function after the calculation
is complete.
</li>
</ul></div>
<p>
Returns:
</p>
<p>
The <code class="computeroutput"><span class="identifier">ArgumentContainer</span></code> corresponding
to the minimum cost found by the optimization.
</p>
<p>
N.B.: The termination criteria is an "OR", not an "AND".
So if the maximum generations is hit, the iteration stops, even if (say) a
<code class="computeroutput"><span class="identifier">value_to_reach</span></code> has not been
attained.
</p>
<h5>
<a name="math_toolkit.differential_evolution.h3"></a>
<span class="phrase"><a name="math_toolkit.differential_evolution.examples"></a></span><a class="link" href="differential_evolution.html#math_toolkit.differential_evolution.examples">Examples</a>
</h5>
<p>
An example use can be found <a href="../../../example/differential_evolution.cpp" target="_top">here</a>.
More examples and API use cases can be studied in <a href="../../../test/differential_evolution_test.cpp" target="_top">differential_evolution_test.cpp</a>.
</p>
<h6>
<a name="math_toolkit.differential_evolution.h4"></a>
<span class="phrase"><a name="math_toolkit.differential_evolution.caveats"></a></span><a class="link" href="differential_evolution.html#math_toolkit.differential_evolution.caveats">Caveats</a>
</h6>
<p>
We have decided to only support classic <code class="computeroutput"><span class="identifier">de</span><span class="special">/</span><span class="identifier">rand</span><span class="special">/</span><span class="number">1</span><span class="special">/</span><span class="identifier">bin</span></code>
because there are too many parameters for this class as it stands, and we have
not seen benchmarks that indicate that other variants of the algorithm perform
better. If a compelling usecase is provided, support for the <code class="computeroutput"><span class="identifier">de</span><span class="special">/</span><span class="identifier">x</span><span class="special">/</span><span class="identifier">y</span><span class="special">/</span><span class="identifier">z</span></code> variants can be added.
</p>
<p>
Supported termination criteria are explicit requests for termination, value-to-reach,
and max generations. Price, Storn, and Lampinen, Section 2.8 also list population
statistics and lack of accepted trials over many generations as sensible termination
criteria. These could be supported if there is demand.
</p>
<p>
Parallelization with <code class="computeroutput"><span class="identifier">std</span><span class="special">::</span><span class="identifier">thread</span></code> does have overhead-especially for
very fast function calls. We found that the function call needs to be roughly
a microsecond for unambigous parallelization wins. However, we have not provided
conditional parallelization as computationally inexpensive cost functions are
the exception; not the rule. If there is a clear use case for conditional parallelization
(cheap cost function in very high dimensions is a good example), we can provide
it.
</p>
<h5>
<a name="math_toolkit.differential_evolution.h5"></a>
<span class="phrase"><a name="math_toolkit.differential_evolution.references"></a></span><a class="link" href="differential_evolution.html#math_toolkit.differential_evolution.references">References</a>
</h5>
<div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; ">
<li class="listitem">
Price, Kenneth, Rainer M. Storn, and Jouni A. Lampinen. <span class="emphasis"><em>Differential
evolution: a practical approach to global optimization.</em></span> Springer
Science &amp; Business Media, 2006.
</li>
<li class="listitem">
D. H. Wolpert and W. G. Macready, <span class="emphasis"><em>No free lunch theorems for
optimization.</em></span> IEEE Transactions on Evolutionary Computation,
vol. 1, no. 1, pp. 67-82, April 1997, doi: 10.1109/4235.585893.
</li>
</ul></div>
</div>
<div class="copyright-footer">Copyright © 2006-2021 Nikhar Agrawal, Anton Bikineev, Matthew Borland,
Paul A. Bristow, Marco Guazzone, Christopher Kormanyos, Hubert Holin, Bruno
Lalande, John Maddock, Evan Miller, Jeremy Murphy, Matthew Pulver, Johan Råde,
Gautam Sewani, Benjamin Sobotta, Nicholas Thompson, Thijs van den Berg, Daryle
Walker and Xiaogang Zhang<p>
Distributed under the Boost Software License, Version 1.0. (See accompanying
file LICENSE_1_0.txt or copy at <a href="http://www.boost.org/LICENSE_1_0.txt" target="_top">http://www.boost.org/LICENSE_1_0.txt</a>)
</p>
</div>
<hr>
<div class="spirit-nav">
<a accesskey="p" href="../optimization.html"><img src="../../../../../doc/src/images/prev.png" alt="Prev"></a><a accesskey="u" href="../optimization.html"><img src="../../../../../doc/src/images/up.png" alt="Up"></a><a accesskey="h" href="../index.html"><img src="../../../../../doc/src/images/home.png" alt="Home"></a><a accesskey="n" href="jso.html"><img src="../../../../../doc/src/images/next.png" alt="Next"></a>
</div>
</body>
</html>
Reference in New Issue View Git Blame Copy Permalink