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.optim.nonlinear.scalar.gradient;
19  
20  /**
21   * This interface represents a preconditioner for differentiable scalar
22   * objective function optimizers.
23   * @since 2.0
24   */
25  public interface Preconditioner {
26      /**
27       * Precondition a search direction.
28       * <p>
29       * The returned preconditioned search direction must be computed fast or
30       * the algorithm performances will drop drastically. A classical approach
31       * is to compute only the diagonal elements of the hessian and to divide
32       * the raw search direction by these elements if they are all positive.
33       * If at least one of them is negative, it is safer to return a clone of
34       * the raw search direction as if the hessian was the identity matrix. The
35       * rationale for this simplified choice is that a negative diagonal element
36       * means the current point is far from the optimum and preconditioning will
37       * not be efficient anyway in this case.
38       * </p>
39       * @param point current point at which the search direction was computed
40       * @param r raw search direction (i.e. opposite of the gradient)
41       * @return approximation of H<sup>-1</sup>r where H is the objective function hessian
42       */
43      double[] precondition(double[] point, double[] r);
44  }