[yul-phaser] Add randomPointCrossover() and fixedPointCrossover() operators

2023-10-03 13:03:40 +00:00 · 2020-02-06 04:34:09 +01:00 · 2020-02-06 04:34:09 +01:00 · c941eaf5d6
commit c941eaf5d6
parent 3fdb4ca607
3 changed files with 238 additions and 0 deletions
--- a/test/yulPhaser/Mutations.cpp
+++ b/test/yulPhaser/Mutations.cpp
@ -212,6 +212,162 @@ BOOST_AUTO_TEST_CASE(alternativeMutations_should_always_choose_second_mutation_i
 		BOOST_TEST(mutation(chromosome) == Chromosome("f"));
 }

+BOOST_AUTO_TEST_CASE(randomPointCrossover_should_swap_chromosome_parts_at_random_point)
+{
+	SimulationRNG::reset(1);
+	function<Crossover> crossover = randomPointCrossover();
+
+	auto [result1, result2] = crossover(Chromosome("aaaaaaaaaa"), Chromosome("cccccc"));
+	BOOST_TEST(result1 == Chromosome("aaaccc"));
+	BOOST_TEST(result2 == Chromosome("cccaaaaaaa"));
+}
+
+BOOST_AUTO_TEST_CASE(randomPointCrossover_should_only_consider_points_available_on_both_chromosomes)
+{
+	SimulationRNG::reset(1);
+	function<Crossover> crossover = randomPointCrossover();
+
+	for (size_t i = 0; i < 30; ++i)
+	{
+		auto [result1, result2] = crossover(Chromosome("aaa"), Chromosome("TTTTTTTTTTTTTTTTTTTT"));
+		BOOST_TEST((
+			(result1 == Chromosome("TTTTTTTTTTTTTTTTTTTT") && result2 == Chromosome("aaa")) ||
+			(result1 == Chromosome("aTTTTTTTTTTTTTTTTTTT") && result2 == Chromosome("Taa")) ||
+			(result1 == Chromosome("aaTTTTTTTTTTTTTTTTTT") && result2 == Chromosome("TTa")) ||
+			(result1 == Chromosome("aaaTTTTTTTTTTTTTTTTT") && result2 == Chromosome("TTT"))
+		));
+	}
+}
+
+BOOST_AUTO_TEST_CASE(randomPointCrossover_should_never_split_at_position_zero_if_chromosomes_are_splittable)
+{
+	SimulationRNG::reset(1);
+	function<Crossover> crossover = randomPointCrossover();
+
+	for (size_t i = 0; i < 30; ++i)
+	{
+		auto [result1, result2] = crossover(Chromosome("aa"), Chromosome("TTTTTTTTTTTTTTTTTTTT"));
+		BOOST_TEST(result1 != Chromosome("TTTTTTTTTTTTTTTTTTTT"));
+		BOOST_TEST(result2 != Chromosome("aa"));
+	}
+}
+
+BOOST_AUTO_TEST_CASE(randomPointCrossover_should_never_split_at_position_zero_if_chromosomes_are_not_empty)
+{
+	SimulationRNG::reset(1);
+	function<Crossover> crossover = randomPointCrossover();
+
+	for (size_t i = 0; i < 30; ++i)
+	{
+		auto [result1, result2] = crossover(Chromosome("a"), Chromosome("T"));
+		BOOST_TEST(result1 == Chromosome("a"));
+		BOOST_TEST(result2 == Chromosome("T"));
+	}
+}
+
+BOOST_AUTO_TEST_CASE(randomPointCrossover_should_work_even_if_one_chromosome_is_unsplittable)
+{
+	function<Crossover> crossover = randomPointCrossover();
+
+	SimulationRNG::reset(1);
+	BOOST_CHECK(crossover(Chromosome("ff"), Chromosome("a")) == ChromosomePair(Chromosome("f"), Chromosome("af")));
+	BOOST_CHECK(crossover(Chromosome("a"), Chromosome("ff")) == ChromosomePair(Chromosome("af"), Chromosome("f")));
+}
+
+BOOST_AUTO_TEST_CASE(randomPointCrossover_should_split_at_position_zero_only_if_at_least_one_chromosome_is_empty)
+{
+	Chromosome empty("");
+	Chromosome unsplittable("a");
+	Chromosome splittable("aaaa");
+	function<Crossover> crossover = randomPointCrossover();
+
+	SimulationRNG::reset(1);
+	BOOST_CHECK(crossover(empty, empty) == ChromosomePair(empty, empty));
+	BOOST_CHECK(crossover(unsplittable, empty) == ChromosomePair(empty, unsplittable));
+	BOOST_CHECK(crossover(empty, unsplittable) == ChromosomePair(unsplittable, empty));
+	BOOST_CHECK(crossover(splittable, empty) == ChromosomePair(empty, splittable));
+	BOOST_CHECK(crossover(empty, splittable) == ChromosomePair(splittable, empty));
+}
+
+BOOST_AUTO_TEST_CASE(fixedPointCrossover_should_swap_chromosome_parts_at_given_point)
+{
+	auto [result1, result2] = fixedPointCrossover(0.8)(Chromosome("aaaaaaaaaa"), Chromosome("cccccccccc"));
+	BOOST_TEST(result1 == Chromosome("aaaaaaaacc"));
+	BOOST_TEST(result2 == Chromosome("ccccccccaa"));
+}
+
+BOOST_AUTO_TEST_CASE(fixedPointCrossover_should_determine_crossover_point_based_on_length_of_shorter_chromosome)
+{
+	auto [result1, result2] = fixedPointCrossover(0.4)(Chromosome("aaaaa"), Chromosome("cccccccccc"));
+	BOOST_TEST(result1 == Chromosome("aacccccccc"));
+	BOOST_TEST(result2 == Chromosome("ccaaa"));
+}
+
+BOOST_AUTO_TEST_CASE(fixedPointCrossover_should_round_split_point)
+{
+	auto [result1, result2] = fixedPointCrossover(0.49)(Chromosome("aaaaa"), Chromosome("ccccc"));
+	BOOST_TEST(result1 == Chromosome("aaccc"));
+	BOOST_TEST(result2 == Chromosome("ccaaa"));
+
+	auto [result3, result4] = fixedPointCrossover(0.50)(Chromosome("aaaaa"), Chromosome("ccccc"));
+	BOOST_TEST(result3 == Chromosome("aaacc"));
+	BOOST_TEST(result4 == Chromosome("cccaa"));
+
+	auto [result5, result6] = fixedPointCrossover(0.51)(Chromosome("aaaaa"), Chromosome("ccccc"));
+	BOOST_TEST(result5 == Chromosome("aaacc"));
+	BOOST_TEST(result6 == Chromosome("cccaa"));
+}
+
+BOOST_AUTO_TEST_CASE(fixedPointCrossover_should_split_at_position_zero_if_explicitly_requested)
+{
+	auto [result1, result2] = fixedPointCrossover(0.0)(Chromosome("aaaaa"), Chromosome("cccccccccc"));
+	BOOST_TEST(result1 == Chromosome("cccccccccc"));
+	BOOST_TEST(result2 == Chromosome("aaaaa"));
+}
+
+BOOST_AUTO_TEST_CASE(fixedPointCrossover_should_split_at_end_of_shorter_chromosome_if_crossover_point_is_after_last_position)
+{
+	auto [result1, result2] = fixedPointCrossover(1.0)(Chromosome("aaaaa"), Chromosome("cccccccccc"));
+	BOOST_TEST(result1 == Chromosome("aaaaaccccc"));
+	BOOST_TEST(result2 == Chromosome("ccccc"));
+}
+
+BOOST_AUTO_TEST_CASE(fixedPointCrossover_should_select_correct_split_point_for_unsplittable_chromosomes)
+{
+	function<Crossover> crossover00 = fixedPointCrossover(0.0);
+	BOOST_CHECK(crossover00(Chromosome("fff"), Chromosome("a")) == ChromosomePair(Chromosome("a"), Chromosome("fff")));
+	BOOST_CHECK(crossover00(Chromosome("a"), Chromosome("fff")) == ChromosomePair(Chromosome("fff"), Chromosome("a")));
+
+	BOOST_CHECK(crossover00(Chromosome("f"), Chromosome("a")) == ChromosomePair(Chromosome("a"), Chromosome("f")));
+
+	function<Crossover> crossover10 = fixedPointCrossover(1.0);
+	BOOST_CHECK(crossover10(Chromosome("fff"), Chromosome("a")) == ChromosomePair(Chromosome("f"), Chromosome("aff")));
+	BOOST_CHECK(crossover10(Chromosome("a"), Chromosome("fff")) == ChromosomePair(Chromosome("aff"), Chromosome("f")));
+
+	BOOST_CHECK(crossover10(Chromosome("f"), Chromosome("a")) == ChromosomePair(Chromosome("f"), Chromosome("a")));
+}
+
+BOOST_AUTO_TEST_CASE(fixedPointCrossover_should_always_use_position_zero_as_split_point_when_chromosome_empty)
+{
+	Chromosome empty("");
+	Chromosome unsplittable("f");
+	Chromosome splittable("aaaa");
+
+	function<Crossover> crossover00 = fixedPointCrossover(0.0);
+	BOOST_CHECK(crossover00(empty, empty) == ChromosomePair(empty, empty));
+	BOOST_CHECK(crossover00(unsplittable, empty) == ChromosomePair(empty, unsplittable));
+	BOOST_CHECK(crossover00(empty, unsplittable) == ChromosomePair(unsplittable, empty));
+	BOOST_CHECK(crossover00(splittable, empty) == ChromosomePair(empty, splittable));
+	BOOST_CHECK(crossover00(empty, splittable) == ChromosomePair(splittable, empty));
+
+	function<Crossover> crossover10 = fixedPointCrossover(1.0);
+	BOOST_CHECK(crossover10(empty, empty) == ChromosomePair(empty, empty));
+	BOOST_CHECK(crossover10(unsplittable, empty) == ChromosomePair(empty, unsplittable));
+	BOOST_CHECK(crossover10(empty, unsplittable) == ChromosomePair(unsplittable, empty));
+	BOOST_CHECK(crossover10(splittable, empty) == ChromosomePair(empty, splittable));
+	BOOST_CHECK(crossover10(empty, splittable) == ChromosomePair(splittable, empty));
+}
+
 BOOST_AUTO_TEST_SUITE_END()
 BOOST_AUTO_TEST_SUITE_END()
 BOOST_AUTO_TEST_SUITE_END()
--- a/tools/yulPhaser/Mutations.cpp
+++ b/tools/yulPhaser/Mutations.cpp
@ -19,6 +19,11 @@

 #include <tools/yulPhaser/SimulationRNG.h>

+#include <libsolutil/CommonData.h>
+
+#include <algorithm>
+#include <cassert>
+#include <cmath>
 #include <string>
 #include <vector>

@ -89,3 +94,60 @@ function<Mutation> phaser::alternativeMutations(
 			return _mutation2(_chromosome);
 	};
 }
+
+namespace
+{
+
+ChromosomePair buildChromosomesBySwappingParts(
+	Chromosome const& _chromosome1,
+	Chromosome const& _chromosome2,
+	size_t _crossoverPoint
+)
+{
+	assert(_crossoverPoint <= _chromosome1.length());
+	assert(_crossoverPoint <= _chromosome2.length());
+
+	auto begin1 = _chromosome1.optimisationSteps().begin();
+	auto begin2 = _chromosome2.optimisationSteps().begin();
+
+	return ChromosomePair(
+		Chromosome(
+			vector<string>(begin1, begin1 + _crossoverPoint) +
+			vector<string>(begin2 + _crossoverPoint, _chromosome2.optimisationSteps().end())
+		),
+		Chromosome(
+			vector<string>(begin2, begin2 + _crossoverPoint) +
+			vector<string>(begin1 + _crossoverPoint, _chromosome1.optimisationSteps().end())
+		)
+	);
+}
+
+}
+
+function<Crossover> phaser::randomPointCrossover()
+{
+	return [=](Chromosome const& _chromosome1, Chromosome const& _chromosome2)
+	{
+		size_t minLength = min(_chromosome1.length(), _chromosome2.length());
+
+		// Don't use position 0 (because this just swaps the values) unless it's the only choice.
+		size_t minPoint = (minLength > 0? 1 : 0);
+		assert(minPoint <= minLength);
+
+		size_t randomPoint = SimulationRNG::uniformInt(minPoint, minLength);
+		return buildChromosomesBySwappingParts(_chromosome1, _chromosome2, randomPoint);
+	};
+}
+
+function<Crossover> phaser::fixedPointCrossover(double _crossoverPoint)
+{
+	assert(0.0 <= _crossoverPoint && _crossoverPoint <= 1.0);
+
+	return [=](Chromosome const& _chromosome1, Chromosome const& _chromosome2)
+	{
+		size_t minLength = min(_chromosome1.length(), _chromosome2.length());
+		size_t concretePoint = static_cast<size_t>(round(minLength * _crossoverPoint));
+
+		return buildChromosomesBySwappingParts(_chromosome1, _chromosome2, concretePoint);
+	};
+}
--- a/tools/yulPhaser/Mutations.h
+++ b/tools/yulPhaser/Mutations.h
@ -23,11 +23,15 @@
 #include <tools/yulPhaser/Chromosome.h>

 #include <functional>
+#include <tuple>

 namespace solidity::phaser
 {

+using ChromosomePair = std::tuple<Chromosome, Chromosome>;
+
 using Mutation = Chromosome(Chromosome const&);
+using Crossover = ChromosomePair(Chromosome const&, Chromosome const&);

 // MUTATIONS

@ -53,4 +57,20 @@ std::function<Mutation> alternativeMutations(
 	std::function<Mutation> _mutation2
 );

+// CROSSOVER
+
+/// Creates a crossover operator that randomly selects a number between 0 and 1 and uses it as the
+/// position at which to perform perform @a fixedPointCrossover.
+std::function<Crossover> randomPointCrossover();
+
+/// Creates a crossover operator that always chooses a point that lies at @a _crossoverPoint
+/// percent of the length of the shorter chromosome. Then creates a pair of chromosomes by
+/// splitting both inputs at the crossover point and stitching the resulting parts. The first
+/// output is created from the first half or first input and the second half of the second input
+/// The second output from the remaining two halves.
+///
+/// Avoids selecting position 0 (since this just produces a chromosome identical to the second one)
+/// unless there is no other choice (i.e. one of the chromosomes is empty).
+std::function<Crossover> fixedPointCrossover(double _crossoverPoint);
+
 }