| 1 | /** \file |
| 2 | * \brief Tests for the A* informed search algorithm. |
| 3 | * |
| 4 | * \author Tilo Wiedera |
| 5 | * |
| 6 | * \par License: |
| 7 | * This file is part of the Open Graph Drawing Framework (OGDF). |
| 8 | * |
| 9 | * \par |
| 10 | * Copyright (C)<br> |
| 11 | * See README.md in the OGDF root directory for details. |
| 12 | * |
| 13 | * \par |
| 14 | * This program is free software; you can redistribute it and/or |
| 15 | * modify it under the terms of the GNU General Public License |
| 16 | * Version 2 or 3 as published by the Free Software Foundation; |
| 17 | * see the file LICENSE.txt included in the packaging of this file |
| 18 | * for details. |
| 19 | * |
| 20 | * \par |
| 21 | * This program is distributed in the hope that it will be useful, |
| 22 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 23 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 24 | * GNU General Public License for more details. |
| 25 | * |
| 26 | * \par |
| 27 | * You should have received a copy of the GNU General Public |
| 28 | * License along with this program; if not, see |
| 29 | * http://www.gnu.org/copyleft/gpl.html |
| 30 | */ |
| 31 | |
| 32 | #include <iomanip> |
| 33 | #include <chrono> |
| 34 | |
| 35 | #include <ogdf/basic/graph_generators.h> |
| 36 | #include <ogdf/graphalg/AStarSearch.h> |
| 37 | #include <ogdf/graphalg/Dijkstra.h> |
| 38 | |
| 39 | #include <testing.h> |
| 40 | |
| 41 | template<typename T> |
| 42 | void validatePath( |
| 43 | const node source, |
| 44 | const node target, |
| 45 | const Graph &graph, |
| 46 | const EdgeArray<T> &cost, |
| 47 | const NodeArray<edge> &pred, |
| 48 | const T expectedCost) { |
| 49 | |
| 50 | T actualCost = 0; |
| 51 | |
| 52 | NodeArray<bool> visited(graph, false); |
| 53 | |
| 54 | for(node v = target; v != source; v = pred[v]->opposite(v)) { |
| 55 | AssertThat(visited[v], IsFalse()); |
| 56 | actualCost += cost[pred[v]]; |
| 57 | visited[v] = true; |
| 58 | } |
| 59 | |
| 60 | AssertThat(actualCost, Equals(expectedCost)); |
| 61 | } |
| 62 | |
| 63 | template<typename T> |
| 64 | void performSingleTest( |
| 65 | const Graph &graph, |
| 66 | const node source, |
| 67 | const node target, |
| 68 | const EdgeArray<T> cost, |
| 69 | const double maxGap, |
| 70 | const bool directed, |
| 71 | Dijkstra<T> &dijkstra, |
| 72 | AStarSearch<T> &astar, |
| 73 | long &ticksDijkstra, |
| 74 | long &ticksUninformedAStar, |
| 75 | long &ticksAStarHeuristic) |
| 76 | { |
| 77 | NodeArray<T> distance(graph, -1); |
| 78 | NodeArray<edge> pred(graph); |
| 79 | |
| 80 | auto start = std::chrono::system_clock::now(); |
| 81 | dijkstra.call(graph, cost, source, pred, distance, directed); |
| 82 | ticksDijkstra += (std::chrono::system_clock::now() - start).count(); |
| 83 | bool foundPath = pred[target] != nullptr; |
| 84 | T opt = distance[target]; |
| 85 | |
| 86 | if(foundPath) { |
| 87 | validatePath(source, target, graph, cost, pred, distance[target]); |
| 88 | |
| 89 | distance.init(graph, -1); |
| 90 | pred.init(graph, nullptr); |
| 91 | |
| 92 | start = std::chrono::system_clock::now(); |
| 93 | T result = astar.call(graph, cost, source, target, pred, [&](node v) { |
| 94 | // utilize distances as returned by Dijkstra for a perfect heuristic |
| 95 | return distance[v]; |
| 96 | }); |
| 97 | ticksAStarHeuristic += (std::chrono::system_clock::now() - start).count(); |
| 98 | |
| 99 | validatePath(source, target, graph, cost, pred, result); |
| 100 | AssertThat(pred[target] != nullptr, IsTrue()); |
| 101 | AssertThat(distance[target], IsLessThan(opt * maxGap + 1)); |
| 102 | } |
| 103 | |
| 104 | start = std::chrono::system_clock::now(); |
| 105 | T result = astar.call(graph, cost, source, target, pred); |
| 106 | ticksUninformedAStar += (std::chrono::system_clock::now() - start).count(); |
| 107 | |
| 108 | AssertThat(pred[target] != nullptr, Equals(foundPath)); |
| 109 | if(foundPath) { |
| 110 | validatePath(source, target, graph, cost, pred, result); |
| 111 | AssertThat(distance[target], IsLessThan(opt * maxGap + 1)); |
| 112 | } |
| 113 | } |
| 114 | |
| 115 | template<typename T> |
| 116 | void performTests(const bool directed, const double maxGap, const bool pathLike) { |
| 117 | const int NUMBER_OF_GRAPHS = 10; |
| 118 | const int MIN_NODES = 100; |
| 119 | const int MAX_NODES = 200; |
| 120 | |
| 121 | AStarSearch<T> astar(directed, maxGap); |
| 122 | Dijkstra<T> dijkstra; |
| 123 | |
| 124 | long ticksDijkstra = 0; |
| 125 | long ticksUninformedAStar = 0; |
| 126 | long ticksAStarHeuristic = 0; |
| 127 | |
| 128 | for(int i = 0; i < NUMBER_OF_GRAPHS; i++) { |
| 129 | Graph graph; |
| 130 | EdgeArray<T> cost(graph); |
| 131 | node source = nullptr; |
| 132 | node target = nullptr; |
| 133 | int n = randomNumber(MIN_NODES, MAX_NODES); |
| 134 | |
| 135 | if(pathLike) { |
| 136 | completeGraph(graph, n); |
| 137 | cost.init(graph, n); |
| 138 | |
| 139 | source = graph.chooseNode(); |
| 140 | node v = source; |
| 141 | |
| 142 | for(int k = 0; k < n/2 || v == source; k++) { |
| 143 | adjEntry adj = v->firstAdj(); |
| 144 | |
| 145 | for(int j = randomNumber(0, v->degree()-1); j > 0; j--) { |
| 146 | adj = adj->succ(); |
| 147 | } |
| 148 | |
| 149 | edge e = adj->theEdge(); |
| 150 | cost[e] = randomNumber(1, 10); |
| 151 | v = e->opposite(v); |
| 152 | } |
| 153 | |
| 154 | target = v; |
| 155 | } else { |
| 156 | randomBiconnectedGraph(graph, n, randomNumber(n, (n*(n-1) / 2))); |
| 157 | |
| 158 | for(edge e : graph.edges) { |
| 159 | cost[e] = randomNumber(1, graph.numberOfEdges()); |
| 160 | } |
| 161 | |
| 162 | source = graph.chooseNode(); |
| 163 | target = graph.chooseNode([&](node v) { return v != source; }); |
| 164 | } |
| 165 | |
| 166 | performSingleTest(graph, source, target, cost, maxGap, directed, dijkstra, astar, |
| 167 | ticksDijkstra, ticksUninformedAStar, ticksAStarHeuristic); |
| 168 | } |
| 169 | |
| 170 | std::cout << std::endl; |
| 171 | std::cout << std::left << " Dijkstra : " << std::right << std::setw(16) << ticksDijkstra << std::endl; |
| 172 | std::cout << std::left << " A* uninformed : " << std::right << std::setw(16) << ticksUninformedAStar << std::endl; |
| 173 | std::cout << std::left << " A* perfect heuristic : " << std::right << std::setw(16) << ticksAStarHeuristic << std::endl; |
| 174 | std::cout << std::left; |
| 175 | } |
| 176 | |
| 177 | template<typename T> |
| 178 | void registerTests(string typeName) { |
| 179 | EpsilonTest et; |
| 180 | for(int i = 0; i < 16; i++) { |
| 181 | bool pathLike = i % 2; |
| 182 | bool directed = (i / 2) % 2; |
| 183 | double maxGap = 1 + (i / 4) / (double) 2; |
| 184 | |
| 185 | string title = "yields the same result as Dijkstra" ; |
| 186 | if(!et.equal(maxGap, 1.0)) { |
| 187 | title = "approximates the optimal solution with a maxmimum gap of " + to_string(maxGap); |
| 188 | } |
| 189 | |
| 190 | title = "[" + typeName + "] " + title; |
| 191 | |
| 192 | title += " on " ; |
| 193 | title += (directed ? "directed " : "" ); |
| 194 | title += (pathLike ? "path-like" : "biconnected" ); |
| 195 | title += " graphs" ; |
| 196 | |
| 197 | it(title, [&](){ |
| 198 | performTests<T>(directed, maxGap, pathLike); |
| 199 | }); |
| 200 | } |
| 201 | } |
| 202 | |
| 203 | go_bandit([](){ |
| 204 | describe("A* Informed Search Algorithm" , [](){ |
| 205 | registerTests<int>("int" ); |
| 206 | registerTests<double>("double" ); |
| 207 | }); |
| 208 | }); |
| 209 | |