001/* 002 * Licensed to the Apache Software Foundation (ASF) under one or more 003 * contributor license agreements. See the NOTICE file distributed with 004 * this work for additional information regarding copyright ownership. 005 * The ASF licenses this file to You under the Apache License, Version 2.0 006 * (the "License"); you may not use this file except in compliance with 007 * the License. You may obtain a copy of the License at 008 * 009 * http://www.apache.org/licenses/LICENSE-2.0 010 * 011 * Unless required by applicable law or agreed to in writing, software 012 * distributed under the License is distributed on an "AS IS" BASIS, 013 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 014 * See the License for the specific language governing permissions and 015 * limitations under the License. 016 */ 017 018package org.apache.commons.math3.ode.nonstiff; 019 020import org.apache.commons.math3.Field; 021import org.apache.commons.math3.RealFieldElement; 022import org.apache.commons.math3.ode.FieldEquationsMapper; 023import org.apache.commons.math3.ode.FieldODEStateAndDerivative; 024import org.apache.commons.math3.util.MathArrays; 025import org.apache.commons.math3.util.MathUtils; 026 027 028/** 029 * This class implements the 5(4) Dormand-Prince integrator for Ordinary 030 * Differential Equations. 031 032 * <p>This integrator is an embedded Runge-Kutta integrator 033 * of order 5(4) used in local extrapolation mode (i.e. the solution 034 * is computed using the high order formula) with stepsize control 035 * (and automatic step initialization) and continuous output. This 036 * method uses 7 functions evaluations per step. However, since this 037 * is an <i>fsal</i>, the last evaluation of one step is the same as 038 * the first evaluation of the next step and hence can be avoided. So 039 * the cost is really 6 functions evaluations per step.</p> 040 * 041 * <p>This method has been published (whithout the continuous output 042 * that was added by Shampine in 1986) in the following article : 043 * <pre> 044 * A family of embedded Runge-Kutta formulae 045 * J. R. Dormand and P. J. Prince 046 * Journal of Computational and Applied Mathematics 047 * volume 6, no 1, 1980, pp. 19-26 048 * </pre></p> 049 * 050 * @param <T> the type of the field elements 051 * @since 3.6 052 */ 053 054public class DormandPrince54FieldIntegrator<T extends RealFieldElement<T>> 055 extends EmbeddedRungeKuttaFieldIntegrator<T> { 056 057 /** Integrator method name. */ 058 private static final String METHOD_NAME = "Dormand-Prince 5(4)"; 059 060 /** Error array, element 1. */ 061 private final T e1; 062 063 // element 2 is zero, so it is neither stored nor used 064 065 /** Error array, element 3. */ 066 private final T e3; 067 068 /** Error array, element 4. */ 069 private final T e4; 070 071 /** Error array, element 5. */ 072 private final T e5; 073 074 /** Error array, element 6. */ 075 private final T e6; 076 077 /** Error array, element 7. */ 078 private final T e7; 079 080 /** Simple constructor. 081 * Build a fifth order Dormand-Prince integrator with the given step bounds 082 * @param field field to which the time and state vector elements belong 083 * @param minStep minimal step (sign is irrelevant, regardless of 084 * integration direction, forward or backward), the last step can 085 * be smaller than this 086 * @param maxStep maximal step (sign is irrelevant, regardless of 087 * integration direction, forward or backward), the last step can 088 * be smaller than this 089 * @param scalAbsoluteTolerance allowed absolute error 090 * @param scalRelativeTolerance allowed relative error 091 */ 092 public DormandPrince54FieldIntegrator(final Field<T> field, 093 final double minStep, final double maxStep, 094 final double scalAbsoluteTolerance, 095 final double scalRelativeTolerance) { 096 super(field, METHOD_NAME, 6, 097 minStep, maxStep, scalAbsoluteTolerance, scalRelativeTolerance); 098 e1 = fraction( 71, 57600); 099 e3 = fraction( -71, 16695); 100 e4 = fraction( 71, 1920); 101 e5 = fraction(-17253, 339200); 102 e6 = fraction( 22, 525); 103 e7 = fraction( -1, 40); 104 } 105 106 /** Simple constructor. 107 * Build a fifth order Dormand-Prince integrator with the given step bounds 108 * @param field field to which the time and state vector elements belong 109 * @param minStep minimal step (sign is irrelevant, regardless of 110 * integration direction, forward or backward), the last step can 111 * be smaller than this 112 * @param maxStep maximal step (sign is irrelevant, regardless of 113 * integration direction, forward or backward), the last step can 114 * be smaller than this 115 * @param vecAbsoluteTolerance allowed absolute error 116 * @param vecRelativeTolerance allowed relative error 117 */ 118 public DormandPrince54FieldIntegrator(final Field<T> field, 119 final double minStep, final double maxStep, 120 final double[] vecAbsoluteTolerance, 121 final double[] vecRelativeTolerance) { 122 super(field, METHOD_NAME, 6, 123 minStep, maxStep, vecAbsoluteTolerance, vecRelativeTolerance); 124 e1 = fraction( 71, 57600); 125 e3 = fraction( -71, 16695); 126 e4 = fraction( 71, 1920); 127 e5 = fraction(-17253, 339200); 128 e6 = fraction( 22, 525); 129 e7 = fraction( -1, 40); 130 } 131 132 /** {@inheritDoc} */ 133 public T[] getC() { 134 final T[] c = MathArrays.buildArray(getField(), 6); 135 c[0] = fraction(1, 5); 136 c[1] = fraction(3, 10); 137 c[2] = fraction(4, 5); 138 c[3] = fraction(8, 9); 139 c[4] = getField().getOne(); 140 c[5] = getField().getOne(); 141 return c; 142 } 143 144 /** {@inheritDoc} */ 145 public T[][] getA() { 146 final T[][] a = MathArrays.buildArray(getField(), 6, -1); 147 for (int i = 0; i < a.length; ++i) { 148 a[i] = MathArrays.buildArray(getField(), i + 1); 149 } 150 a[0][0] = fraction( 1, 5); 151 a[1][0] = fraction( 3, 40); 152 a[1][1] = fraction( 9, 40); 153 a[2][0] = fraction( 44, 45); 154 a[2][1] = fraction( -56, 15); 155 a[2][2] = fraction( 32, 9); 156 a[3][0] = fraction( 19372, 6561); 157 a[3][1] = fraction(-25360, 2187); 158 a[3][2] = fraction( 64448, 6561); 159 a[3][3] = fraction( -212, 729); 160 a[4][0] = fraction( 9017, 3168); 161 a[4][1] = fraction( -355, 33); 162 a[4][2] = fraction( 46732, 5247); 163 a[4][3] = fraction( 49, 176); 164 a[4][4] = fraction( -5103, 18656); 165 a[5][0] = fraction( 35, 384); 166 a[5][1] = getField().getZero(); 167 a[5][2] = fraction( 500, 1113); 168 a[5][3] = fraction( 125, 192); 169 a[5][4] = fraction( -2187, 6784); 170 a[5][5] = fraction( 11, 84); 171 return a; 172 } 173 174 /** {@inheritDoc} */ 175 public T[] getB() { 176 final T[] b = MathArrays.buildArray(getField(), 7); 177 b[0] = fraction( 35, 384); 178 b[1] = getField().getZero(); 179 b[2] = fraction( 500, 1113); 180 b[3] = fraction( 125, 192); 181 b[4] = fraction(-2187, 6784); 182 b[5] = fraction( 11, 84); 183 b[6] = getField().getZero(); 184 return b; 185 } 186 187 /** {@inheritDoc} */ 188 @Override 189 protected DormandPrince54FieldStepInterpolator<T> 190 createInterpolator(final boolean forward, T[][] yDotK, 191 final FieldODEStateAndDerivative<T> globalPreviousState, 192 final FieldODEStateAndDerivative<T> globalCurrentState, final FieldEquationsMapper<T> mapper) { 193 return new DormandPrince54FieldStepInterpolator<T>(getField(), forward, yDotK, 194 globalPreviousState, globalCurrentState, 195 globalPreviousState, globalCurrentState, 196 mapper); 197 } 198 199 /** {@inheritDoc} */ 200 @Override 201 public int getOrder() { 202 return 5; 203 } 204 205 /** {@inheritDoc} */ 206 @Override 207 protected T estimateError(final T[][] yDotK, final T[] y0, final T[] y1, final T h) { 208 209 T error = getField().getZero(); 210 211 for (int j = 0; j < mainSetDimension; ++j) { 212 final T errSum = yDotK[0][j].multiply(e1). 213 add(yDotK[2][j].multiply(e3)). 214 add(yDotK[3][j].multiply(e4)). 215 add(yDotK[4][j].multiply(e5)). 216 add(yDotK[5][j].multiply(e6)). 217 add(yDotK[6][j].multiply(e7)); 218 219 final T yScale = MathUtils.max(y0[j].abs(), y1[j].abs()); 220 final T tol = (vecAbsoluteTolerance == null) ? 221 yScale.multiply(scalRelativeTolerance).add(scalAbsoluteTolerance) : 222 yScale.multiply(vecRelativeTolerance[j]).add(vecAbsoluteTolerance[j]); 223 final T ratio = h.multiply(errSum).divide(tol); 224 error = error.add(ratio.multiply(ratio)); 225 226 } 227 228 return error.divide(mainSetDimension).sqrt(); 229 230 } 231 232}