本文整理汇总了Java中org.apache.commons.math3.analysis.solvers.BrentSolver.solve方法的典型用法代码示例。如果您正苦于以下问题：Java BrentSolver.solve方法的具体用法？Java BrentSolver.solve怎么用？Java BrentSolver.solve使用的例子？那么恭喜您, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类org.apache.commons.math3.analysis.solvers.BrentSolver的用法示例。
在下文中一共展示了BrentSolver.solve方法的3个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Java代码示例。
示例1: testSolveWaterVapourFunction
import org.apache.commons.math3.analysis.solvers.BrentSolver; //导入方法依赖的package包/类
@Test
public void testSolveWaterVapourFunction() throws Exception {
// all numbers in this test taken from IDL test run, cellIndexX=1, cellIndexY=0
WaterVapourFunction wvFunction = new WaterVapourFunction();
wvFunction.setMerisRatio(0.64714217);
wvFunction.setWvGr2(new double[]{0.301, 1.0, 1.5, 2.0, 2.7, 4.999});
final double[][][] parAtmH = initParAtmH();
wvFunction.setParAtmH(parAtmH);
wvFunction.setReflPix(new double[]{0.358543, 0.365387});
// Define the fractional tolerance:
final double ftol = 1.0e-4;
final int maxIter = 10000;
final double wvLower = 0.302;
final double wvUpper = 4.998;
BrentSolver brentSolver = new BrentSolver(ftol);
double result = brentSolver.solve(maxIter, wvFunction, wvLower, wvUpper);
assertEquals(1.96476, result, ftol);
}
示例2: testDekkerBrentFunction
import org.apache.commons.math3.analysis.solvers.BrentSolver; //导入方法依赖的package包/类
@Test
public void testDekkerBrentFunction() throws Exception {
// Define the fractional tolerance:
final double ftol = 1.0e-4;
final int maxIter = 10000;
// f(x) = x - 1
DekkerBrentTestFunction1 testFunction = new DekkerBrentTestFunction1();
double a = 0.5;
double b = 2.0;
BrentSolver brentSolver = new BrentSolver(ftol);
double result = brentSolver.solve(maxIter, testFunction, a, b);
assertEquals(1.0, result, 1.E-4);
// f(x) = cos(x)
DekkerBrentTestFunction2 testFunction2 = new DekkerBrentTestFunction2();
a = 1.0;
b = 2.0;
result = brentSolver.solve(maxIter, testFunction2, a, b);
assertEquals(Math.PI/2.0, result, 1.E-4);
}
示例3: simpleTest
import org.apache.commons.math3.analysis.solvers.BrentSolver; //导入方法依赖的package包/类
/**
* Test on a 4th degree polynomial with 4 real roots at x = 0, 1, 2, 3. This objective function is positive for
* large enough positive and negative values of its arguments. Therefore, the simple Brent solver complains that
* the search interval does not bracket a root. The robust Brent solver, however, subdivides the given search
* interval and finds a bracketing sub-interval.
*
* The "best" root according to the given merit function (set to the anti-derivative of the objective function)
* is in fact the one at x = 0. We require the robust solver to output x = 0, and the simple solver to fail.
*/
@Test
public void simpleTest() {
final UnivariateFunction objFunc = x -> 30 * x * (x - 1) * (x - 2) * (x - 3);
final UnivariateFunction meritFunc = x -> 6 * FastMath.pow(x, 5) - 45 * FastMath.pow(x, 4) + 110 * FastMath.pow(x, 3) -
90 * FastMath.pow(x, 2);
final RobustBrentSolver solverRobust = new RobustBrentSolver(DEF_REL_ACC, DEF_REL_ACC, DEF_F_ACC,
meritFunc, 4, 1);
final BrentSolver solverSimple = new BrentSolver(DEF_REL_ACC, DEF_REL_ACC, DEF_F_ACC);
final double xRobust = solverRobust.solve(100, objFunc, -1, 4);
Assert.assertEquals(xRobust, 0, DEF_ABS_ACC);
boolean simpleSolverFails = false;
try {
/* this will fail */
solverSimple.solve(100, objFunc, -1, 4);
} catch (final NoBracketingException ex) {
simpleSolverFails = true;
}
Assert.assertTrue(simpleSolverFails);
}