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    
018    package org.apache.commons.math3.optim.nonlinear.scalar.gradient;
019    
020    /**
021     * This interface represents a preconditioner for differentiable scalar
022     * objective function optimizers.
023     * @version $Id: Preconditioner.java 1416643 2012-12-03 19:37:14Z tn $
024     * @since 2.0
025     */
026    public interface Preconditioner {
027        /**
028         * Precondition a search direction.
029         * <p>
030         * The returned preconditioned search direction must be computed fast or
031         * the algorithm performances will drop drastically. A classical approach
032         * is to compute only the diagonal elements of the hessian and to divide
033         * the raw search direction by these elements if they are all positive.
034         * If at least one of them is negative, it is safer to return a clone of
035         * the raw search direction as if the hessian was the identity matrix. The
036         * rationale for this simplified choice is that a negative diagonal element
037         * means the current point is far from the optimum and preconditioning will
038         * not be efficient anyway in this case.
039         * </p>
040         * @param point current point at which the search direction was computed
041         * @param r raw search direction (i.e. opposite of the gradient)
042         * @return approximation of H<sup>-1</sup>r where H is the objective function hessian
043         */
044        double[] precondition(double[] point, double[] r);
045    }