LocalSearch Class Reference
The super class for local search methods. More...
#include <LocalSearch.h>
Inheritance diagram for LocalSearch:
Public Member Functions | |
| LocalSearch (ObjectiveFunction *objectiveFunction) | |
| Constructor. | |
| ~LocalSearch () | |
| Destructor. | |
| virtual double | search (vector< double > &x) |
| Actual search strategy. | |
Protected Member Functions | |
| bool | done () |
| Check if stopping criteria are satisfied. | |
| double | evaluate (vector< double > &x) |
| Evaluate the objective function. | |
Detailed Description
The super class for local search methods.<File comment goes here!!>
Copyright (c) 2008 by <Quang Huy / NTU>
- Remarks:
- Note that the nature of global search is to minimize due to the search strategy generally based on gradients / hessian matrix / quadratic approximation. Therefore, if the objective function is to maximize, the evaluation should return the negative value of it.
- See also:
- ObjectiveFunction
Constructor & Destructor Documentation
| LocalSearch::LocalSearch | ( | ObjectiveFunction * | objectiveFunction | ) |
Constructor.
<File comment goes here!!>
Copyright (c) 2008 by <Quang Huy / NTU>
- Parameters:
-
objectiveFunction Objective function to be optimized.
- Remarks:
- Note that the nature of global search is to minimize due to the search strategy generally based on gradients / hessian matrix / quadratic approximation. Therefore, if the objective function is to maximize, the evaluation should return the negative value of it.
- See also:
- ObjectiveFunction
Member Function Documentation
| bool LocalSearch::done | ( | ) | [protected] |
Check if stopping criteria are satisfied.
- Returns:
- True if one of the stopping conditions is true, false otherwise.
| double LocalSearch::evaluate | ( | vector< double > & | x | ) | [protected] |
Evaluate the objective function.
- Parameters:
-
x Input phenotype.
- Returns:
- Fitness value.
- Remarks:
- Note that the nature of global search is to minimize due to the search strategy generally based on gradients / hessian matrix / quadratic approximation. Therefore, if the objective function is to maximize, the evaluation should return the negative value of it.
- See also:
- ObjectiveFunction
| double LocalSearch::search | ( | vector< double > & | x | ) | [virtual] |
Actual search strategy.
- Parameters:
-
x Initial point.
- Returns:
- Final finess value.
- Remarks:
- To be inherited by different search strategies.
- See also:
- LocalSearch_DFP | LocalSearch_DSCG
Reimplemented in LocalSearch_DFP, and LocalSearch_DSCG.
The documentation for this class was generated from the following files:
- D:/HUYNQ/Works/Codes/Dolphin/Dolphin/LocalSearches/LocalSearch.h
- D:/HUYNQ/Works/Codes/Dolphin/Dolphin/LocalSearches/LocalSearch.cpp
