Class strategy::LinearSearchOpt
Linear Search Global Optimization stategy searching implementation. Implemented 1 dimensional search and refine algorithm. Solution is computed at sample points iteratively and search interval defined by sample points will be moved toward the end where the best solution is observed.
Inherited Members
Constructors
LinearSearchOpt()
Declaration
strategy::LinearSearchOpt::LinearSearchOpt()
LinearSearchOpt()
Declaration
strategy::LinearSearchOpt::LinearSearchOpt(const LinearSearchOpt&)=delete
Methods
operator=()
Declaration
LinearSearchOpt&strategy::LinearSearchOpt::operator=(const LinearSearchOpt&)=delete
configure()
Declaration
void strategy::LinearSearchOpt::configure(const utils::Json¶ms, int thread_count)
init()
Declaration
void strategy::LinearSearchOpt::init(size_t dimensions, uint32_t seed, size_t reserved_samples=128)
recommend_parameter_values()
Declaration
bool strategy::LinearSearchOpt::recommend_parameter_values(std::vector<double>¶meters_new)
add_new_sample()
Declaration
void strategy::LinearSearchOpt::add_new_sample(std::vector<double>¶meters, double objective)
set_ranges()
Declaration
void strategy::LinearSearchOpt::set_ranges(const std::vector<int>&intial_sample_points)
fit_sample_points()
Search interval defined by sample points will be stretched if the best solution is produced at the last sample point. Search interval defined by sample points will be shrinked if the best solution is produced at the first sample point.
Declaration
void strategy::LinearSearchOpt::fit_sample_points()