View Javadoc
1   /*
2    * Licensed to the Apache Software Foundation (ASF) under one or more
3    * contributor license agreements.  See the NOTICE file distributed with
4    * this work for additional information regarding copyright ownership.
5    * The ASF licenses this file to You under the Apache License, Version 2.0
6    * (the "License"); you may not use this file except in compliance with
7    * the License.  You may obtain a copy of the License at
8    *
9    *      http://www.apache.org/licenses/LICENSE-2.0
10   *
11   * Unless required by applicable law or agreed to in writing, software
12   * distributed under the License is distributed on an "AS IS" BASIS,
13   * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
14   * See the License for the specific language governing permissions and
15   * limitations under the License.
16   */
17  
18  package org.apache.commons.math4.legacy.ode.nonstiff;
19  
20  import java.util.Arrays;
21  
22  import org.apache.commons.math4.legacy.exception.DimensionMismatchException;
23  import org.apache.commons.math4.legacy.exception.MaxCountExceededException;
24  import org.apache.commons.math4.legacy.exception.NoBracketingException;
25  import org.apache.commons.math4.legacy.exception.NumberIsTooSmallException;
26  import org.apache.commons.math4.legacy.linear.Array2DRowRealMatrix;
27  import org.apache.commons.math4.legacy.linear.RealMatrixPreservingVisitor;
28  import org.apache.commons.math4.legacy.ode.EquationsMapper;
29  import org.apache.commons.math4.legacy.ode.ExpandableStatefulODE;
30  import org.apache.commons.math4.legacy.ode.sampling.NordsieckStepInterpolator;
31  import org.apache.commons.math4.core.jdkmath.JdkMath;
32  
33  
34  /**
35   * This class implements implicit Adams-Moulton integrators for Ordinary
36   * Differential Equations.
37   *
38   * <p>Adams-Moulton methods (in fact due to Adams alone) are implicit
39   * multistep ODE solvers. This implementation is a variation of the classical
40   * one: it uses adaptive stepsize to implement error control, whereas
41   * classical implementations are fixed step size. The value of state vector
42   * at step n+1 is a simple combination of the value at step n and of the
43   * derivatives at steps n+1, n, n-1 ... Since y'<sub>n+1</sub> is needed to
44   * compute y<sub>n+1</sub>, another method must be used to compute a first
45   * estimate of y<sub>n+1</sub>, then compute y'<sub>n+1</sub>, then compute
46   * a final estimate of y<sub>n+1</sub> using the following formulas. Depending
47   * on the number k of previous steps one wants to use for computing the next
48   * value, different formulas are available for the final estimate:</p>
49   * <ul>
50   *   <li>k = 1: y<sub>n+1</sub> = y<sub>n</sub> + h y'<sub>n+1</sub></li>
51   *   <li>k = 2: y<sub>n+1</sub> = y<sub>n</sub> + h (y'<sub>n+1</sub>+y'<sub>n</sub>)/2</li>
52   *   <li>k = 3: y<sub>n+1</sub> = y<sub>n</sub> + h (5y'<sub>n+1</sub>+8y'<sub>n</sub>-y'<sub>n-1</sub>)/12</li>
53   *   <li>k = 4: y<sub>n+1</sub> = y<sub>n</sub> + h (9y'<sub>n+1</sub>+19y'<sub>n</sub>-5y'<sub>n-1</sub>+y'<sub>n-2</sub>)/24</li>
54   *   <li>...</li>
55   * </ul>
56   *
57   * <p>A k-steps Adams-Moulton method is of order k+1.</p>
58   *
59   * <p><b>Implementation details</b></p>
60   *
61   * <p>We define scaled derivatives s<sub>i</sub>(n) at step n as:
62   * <div style="white-space: pre"><code>
63   * s<sub>1</sub>(n) = h y'<sub>n</sub> for first derivative
64   * s<sub>2</sub>(n) = h<sup>2</sup>/2 y''<sub>n</sub> for second derivative
65   * s<sub>3</sub>(n) = h<sup>3</sup>/6 y'''<sub>n</sub> for third derivative
66   * ...
67   * s<sub>k</sub>(n) = h<sup>k</sup>/k! y<sup>(k)</sup><sub>n</sub> for k<sup>th</sup> derivative
68   * </code></div>
69   *
70   * <p>The definitions above use the classical representation with several previous first
71   * derivatives. Lets define
72   * <div style="white-space: pre"><code>
73   *   q<sub>n</sub> = [ s<sub>1</sub>(n-1) s<sub>1</sub>(n-2) ... s<sub>1</sub>(n-(k-1)) ]<sup>T</sup>
74   * </code></div>
75   * (we omit the k index in the notation for clarity). With these definitions,
76   * Adams-Moulton methods can be written:
77   * <ul>
78   *   <li>k = 1: y<sub>n+1</sub> = y<sub>n</sub> + s<sub>1</sub>(n+1)</li>
79   *   <li>k = 2: y<sub>n+1</sub> = y<sub>n</sub> + 1/2 s<sub>1</sub>(n+1) + [ 1/2 ] q<sub>n+1</sub></li>
80   *   <li>k = 3: y<sub>n+1</sub> = y<sub>n</sub> + 5/12 s<sub>1</sub>(n+1) + [ 8/12 -1/12 ] q<sub>n+1</sub></li>
81   *   <li>k = 4: y<sub>n+1</sub> = y<sub>n</sub> + 9/24 s<sub>1</sub>(n+1) + [ 19/24 -5/24 1/24 ] q<sub>n+1</sub></li>
82   *   <li>...</li>
83   * </ul>
84   *
85   * <p>Instead of using the classical representation with first derivatives only (y<sub>n</sub>,
86   * s<sub>1</sub>(n+1) and q<sub>n+1</sub>), our implementation uses the Nordsieck vector with
87   * higher degrees scaled derivatives all taken at the same step (y<sub>n</sub>, s<sub>1</sub>(n)
88   * and r<sub>n</sub>) where r<sub>n</sub> is defined as:
89   * <div style="white-space: pre"><code>
90   * r<sub>n</sub> = [ s<sub>2</sub>(n), s<sub>3</sub>(n) ... s<sub>k</sub>(n) ]<sup>T</sup>
91   * </code></div>
92   * (here again we omit the k index in the notation for clarity)
93   *
94   * <p>Taylor series formulas show that for any index offset i, s<sub>1</sub>(n-i) can be
95   * computed from s<sub>1</sub>(n), s<sub>2</sub>(n) ... s<sub>k</sub>(n), the formula being exact
96   * for degree k polynomials.
97   * <div style="white-space: pre"><code>
98   * s<sub>1</sub>(n-i) = s<sub>1</sub>(n) + &sum;<sub>j&gt;0</sub> (j+1) (-i)<sup>j</sup> s<sub>j+1</sub>(n)
99   * </code></div>
100  * The previous formula can be used with several values for i to compute the transform between
101  * classical representation and Nordsieck vector. The transform between r<sub>n</sub>
102  * and q<sub>n</sub> resulting from the Taylor series formulas above is:
103  * <div style="white-space: pre"><code>
104  * q<sub>n</sub> = s<sub>1</sub>(n) u + P r<sub>n</sub>
105  * </code></div>
106  * where u is the [ 1 1 ... 1 ]<sup>T</sup> vector and P is the (k-1)&times;(k-1) matrix built
107  * with the (j+1) (-i)<sup>j</sup> terms with i being the row number starting from 1 and j being
108  * the column number starting from 1:
109  * <pre>
110  *        [  -2   3   -4    5  ... ]
111  *        [  -4  12  -32   80  ... ]
112  *   P =  [  -6  27 -108  405  ... ]
113  *        [  -8  48 -256 1280  ... ]
114  *        [          ...           ]
115  * </pre>
116  *
117  * <p>Using the Nordsieck vector has several advantages:
118  * <ul>
119  *   <li>it greatly simplifies step interpolation as the interpolator mainly applies
120  *   Taylor series formulas,</li>
121  *   <li>it simplifies step changes that occur when discrete events that truncate
122  *   the step are triggered,</li>
123  *   <li>it allows to extend the methods in order to support adaptive stepsize.</li>
124  * </ul>
125  *
126  * <p>The predicted Nordsieck vector at step n+1 is computed from the Nordsieck vector at step
127  * n as follows:
128  * <ul>
129  *   <li>Y<sub>n+1</sub> = y<sub>n</sub> + s<sub>1</sub>(n) + u<sup>T</sup> r<sub>n</sub></li>
130  *   <li>S<sub>1</sub>(n+1) = h f(t<sub>n+1</sub>, Y<sub>n+1</sub>)</li>
131  *   <li>R<sub>n+1</sub> = (s<sub>1</sub>(n) - S<sub>1</sub>(n+1)) P<sup>-1</sup> u + P<sup>-1</sup> A P r<sub>n</sub></li>
132  * </ul>
133  * where A is a rows shifting matrix (the lower left part is an identity matrix):
134  * <pre>
135  *        [ 0 0   ...  0 0 | 0 ]
136  *        [ ---------------+---]
137  *        [ 1 0   ...  0 0 | 0 ]
138  *    A = [ 0 1   ...  0 0 | 0 ]
139  *        [       ...      | 0 ]
140  *        [ 0 0   ...  1 0 | 0 ]
141  *        [ 0 0   ...  0 1 | 0 ]
142  * </pre>
143  * From this predicted vector, the corrected vector is computed as follows:
144  * <ul>
145  *   <li>y<sub>n+1</sub> = y<sub>n</sub> + S<sub>1</sub>(n+1) + [ -1 +1 -1 +1 ... &plusmn;1 ] r<sub>n+1</sub></li>
146  *   <li>s<sub>1</sub>(n+1) = h f(t<sub>n+1</sub>, y<sub>n+1</sub>)</li>
147  *   <li>r<sub>n+1</sub> = R<sub>n+1</sub> + (s<sub>1</sub>(n+1) - S<sub>1</sub>(n+1)) P<sup>-1</sup> u</li>
148  * </ul>
149  * where the upper case Y<sub>n+1</sub>, S<sub>1</sub>(n+1) and R<sub>n+1</sub> represent the
150  * predicted states whereas the lower case y<sub>n+1</sub>, s<sub>n+1</sub> and r<sub>n+1</sub>
151  * represent the corrected states.
152  *
153  * <p>The P<sup>-1</sup>u vector and the P<sup>-1</sup> A P matrix do not depend on the state,
154  * they only depend on k and therefore are precomputed once for all.</p>
155  *
156  * @since 2.0
157  */
158 public class AdamsMoultonIntegrator extends AdamsIntegrator {
159 
160     /** Integrator method name. */
161     private static final String METHOD_NAME = "Adams-Moulton";
162 
163     /**
164      * Build an Adams-Moulton integrator with the given order and error control parameters.
165      * @param nSteps number of steps of the method excluding the one being computed
166      * @param minStep minimal step (sign is irrelevant, regardless of
167      * integration direction, forward or backward), the last step can
168      * be smaller than this
169      * @param maxStep maximal step (sign is irrelevant, regardless of
170      * integration direction, forward or backward), the last step can
171      * be smaller than this
172      * @param scalAbsoluteTolerance allowed absolute error
173      * @param scalRelativeTolerance allowed relative error
174      * @exception NumberIsTooSmallException if order is 1 or less
175      */
176     public AdamsMoultonIntegrator(final int nSteps,
177                                   final double minStep, final double maxStep,
178                                   final double scalAbsoluteTolerance,
179                                   final double scalRelativeTolerance)
180         throws NumberIsTooSmallException {
181         super(METHOD_NAME, nSteps, nSteps + 1, minStep, maxStep,
182               scalAbsoluteTolerance, scalRelativeTolerance);
183     }
184 
185     /**
186      * Build an Adams-Moulton integrator with the given order and error control parameters.
187      * @param nSteps number of steps of the method excluding the one being computed
188      * @param minStep minimal step (sign is irrelevant, regardless of
189      * integration direction, forward or backward), the last step can
190      * be smaller than this
191      * @param maxStep maximal step (sign is irrelevant, regardless of
192      * integration direction, forward or backward), the last step can
193      * be smaller than this
194      * @param vecAbsoluteTolerance allowed absolute error
195      * @param vecRelativeTolerance allowed relative error
196      * @exception IllegalArgumentException if order is 1 or less
197      */
198     public AdamsMoultonIntegrator(final int nSteps,
199                                   final double minStep, final double maxStep,
200                                   final double[] vecAbsoluteTolerance,
201                                   final double[] vecRelativeTolerance)
202         throws IllegalArgumentException {
203         super(METHOD_NAME, nSteps, nSteps + 1, minStep, maxStep,
204               vecAbsoluteTolerance, vecRelativeTolerance);
205     }
206 
207     /** {@inheritDoc} */
208     @Override
209     public void integrate(final ExpandableStatefulODE equations,final double t)
210         throws NumberIsTooSmallException, DimensionMismatchException,
211                MaxCountExceededException, NoBracketingException {
212 
213         sanityChecks(equations, t);
214         setEquations(equations);
215         final boolean forward = t > equations.getTime();
216 
217         // initialize working arrays
218         final double[] y0   = equations.getCompleteState();
219         final double[] y    = y0.clone();
220         final double[] yDot = new double[y.length];
221         final double[] yTmp = new double[y.length];
222         final double[] predictedScaled = new double[y.length];
223         Array2DRowRealMatrix nordsieckTmp = null;
224 
225         // set up two interpolators sharing the integrator arrays
226         final NordsieckStepInterpolator interpolator = new NordsieckStepInterpolator();
227         interpolator.reinitialize(y, forward,
228                                   equations.getPrimaryMapper(), equations.getSecondaryMappers());
229 
230         // set up integration control objects
231         initIntegration(equations.getTime(), y0, t);
232 
233         // compute the initial Nordsieck vector using the configured starter integrator
234         start(equations.getTime(), y, t);
235         interpolator.reinitialize(stepStart, stepSize, scaled, nordsieck);
236         interpolator.storeTime(stepStart);
237 
238         double hNew = stepSize;
239         interpolator.rescale(hNew);
240 
241         isLastStep = false;
242         do {
243 
244             double error = 10;
245             while (error >= 1.0) {
246 
247                 stepSize = hNew;
248 
249                 // predict a first estimate of the state at step end (P in the PECE sequence)
250                 final double stepEnd = stepStart + stepSize;
251                 interpolator.setInterpolatedTime(stepEnd);
252                 final ExpandableStatefulODE expandable = getExpandable();
253                 final EquationsMapper primary = expandable.getPrimaryMapper();
254                 primary.insertEquationData(interpolator.getInterpolatedState(), yTmp);
255                 int index = 0;
256                 for (final EquationsMapper secondary : expandable.getSecondaryMappers()) {
257                     secondary.insertEquationData(interpolator.getInterpolatedSecondaryState(index), yTmp);
258                     ++index;
259                 }
260 
261                 // evaluate a first estimate of the derivative (first E in the PECE sequence)
262                 computeDerivatives(stepEnd, yTmp, yDot);
263 
264                 // update Nordsieck vector
265                 for (int j = 0; j < y0.length; ++j) {
266                     predictedScaled[j] = stepSize * yDot[j];
267                 }
268                 nordsieckTmp = updateHighOrderDerivativesPhase1(nordsieck);
269                 updateHighOrderDerivativesPhase2(scaled, predictedScaled, nordsieckTmp);
270 
271                 // apply correction (C in the PECE sequence)
272                 error = nordsieckTmp.walkInOptimizedOrder(new Corrector(y, predictedScaled, yTmp));
273 
274                 if (error >= 1.0) {
275                     // reject the step and attempt to reduce error by stepsize control
276                     final double factor = computeStepGrowShrinkFactor(error);
277                     hNew = filterStep(stepSize * factor, forward, false);
278                     interpolator.rescale(hNew);
279                 }
280             }
281 
282             // evaluate a final estimate of the derivative (second E in the PECE sequence)
283             final double stepEnd = stepStart + stepSize;
284             computeDerivatives(stepEnd, yTmp, yDot);
285 
286             // update Nordsieck vector
287             final double[] correctedScaled = new double[y0.length];
288             for (int j = 0; j < y0.length; ++j) {
289                 correctedScaled[j] = stepSize * yDot[j];
290             }
291             updateHighOrderDerivativesPhase2(predictedScaled, correctedScaled, nordsieckTmp);
292 
293             // discrete events handling
294             System.arraycopy(yTmp, 0, y, 0, y.length);
295             interpolator.reinitialize(stepEnd, stepSize, correctedScaled, nordsieckTmp);
296             interpolator.storeTime(stepStart);
297             interpolator.shift();
298             interpolator.storeTime(stepEnd);
299             stepStart = acceptStep(interpolator, y, yDot, t);
300             scaled    = correctedScaled;
301             nordsieck = nordsieckTmp;
302 
303             if (!isLastStep) {
304 
305                 // prepare next step
306                 interpolator.storeTime(stepStart);
307 
308                 if (resetOccurred) {
309                     // some events handler has triggered changes that
310                     // invalidate the derivatives, we need to restart from scratch
311                     start(stepStart, y, t);
312                     interpolator.reinitialize(stepStart, stepSize, scaled, nordsieck);
313                 }
314 
315                 // stepsize control for next step
316                 final double  factor     = computeStepGrowShrinkFactor(error);
317                 final double  scaledH    = stepSize * factor;
318                 final double  nextT      = stepStart + scaledH;
319                 final boolean nextIsLast = forward ? (nextT >= t) : (nextT <= t);
320                 hNew = filterStep(scaledH, forward, nextIsLast);
321 
322                 final double  filteredNextT      = stepStart + hNew;
323                 final boolean filteredNextIsLast = forward ? (filteredNextT >= t) : (filteredNextT <= t);
324                 if (filteredNextIsLast) {
325                     hNew = t - stepStart;
326                 }
327 
328                 interpolator.rescale(hNew);
329             }
330         } while (!isLastStep);
331 
332         // dispatch results
333         equations.setTime(stepStart);
334         equations.setCompleteState(y);
335 
336         resetInternalState();
337     }
338 
339     /** Corrector for current state in Adams-Moulton method.
340      * <p>
341      * This visitor implements the Taylor series formula:
342      * <pre>
343      * Y<sub>n+1</sub> = y<sub>n</sub> + s<sub>1</sub>(n+1) + [ -1 +1 -1 +1 ... &plusmn;1 ] r<sub>n+1</sub>
344      * </pre>
345      * </p>
346      */
347     private final class Corrector implements RealMatrixPreservingVisitor {
348 
349         /** Previous state. */
350         private final double[] previous;
351 
352         /** Current scaled first derivative. */
353         private final double[] scaled;
354 
355         /** Current state before correction. */
356         private final double[] before;
357 
358         /** Current state after correction. */
359         private final double[] after;
360 
361         /** Simple constructor.
362          * @param previous previous state
363          * @param scaled current scaled first derivative
364          * @param state state to correct (will be overwritten after visit)
365          */
366         Corrector(final double[] previous, final double[] scaled, final double[] state) {
367             this.previous = previous;
368             this.scaled   = scaled;
369             this.after    = state;
370             this.before   = state.clone();
371         }
372 
373         /** {@inheritDoc} */
374         @Override
375         public void start(int rows, int columns,
376                           int startRow, int endRow, int startColumn, int endColumn) {
377             Arrays.fill(after, 0.0);
378         }
379 
380         /** {@inheritDoc} */
381         @Override
382         public void visit(int row, int column, double value) {
383             if ((row & 0x1) == 0) {
384                 after[column] -= value;
385             } else {
386                 after[column] += value;
387             }
388         }
389 
390         /**
391          * End visiting the Nordsieck vector.
392          * <p>The correction is used to control stepsize. So its amplitude is
393          * considered to be an error, which must be normalized according to
394          * error control settings. If the normalized value is greater than 1,
395          * the correction was too large and the step must be rejected.</p>
396          * @return the normalized correction, if greater than 1, the step
397          * must be rejected
398          */
399         @Override
400         public double end() {
401 
402             double error = 0;
403             for (int i = 0; i < after.length; ++i) {
404                 after[i] += previous[i] + scaled[i];
405                 if (i < mainSetDimension) {
406                     final double yScale = JdkMath.max(JdkMath.abs(previous[i]), JdkMath.abs(after[i]));
407                     final double tol    = (vecAbsoluteTolerance == null) ?
408                                           (scalAbsoluteTolerance + scalRelativeTolerance * yScale) :
409                                           (vecAbsoluteTolerance[i] + vecRelativeTolerance[i] * yScale);
410                     final double ratio  = (after[i] - before[i]) / tol; // (corrected-predicted)/tol
411                     error += ratio * ratio;
412                 }
413             }
414 
415             return JdkMath.sqrt(error / mainSetDimension);
416         }
417     }
418 }