Struct minimax::strategies::iterative::IterativeOptions

pub struct IterativeOptions {
    pub table_byte_size: usize,
    pub verbose: bool,
    /* private fields */
}

Options to use for the iterative search engines.

Fields

table_byte_size: usize

verbose: bool

Implementations

impl IterativeOptions

pub fn new() -> Self

impl IterativeOptions

pub fn with_table_byte_size(self, size: usize) -> Self

Approximately how large the transposition table should be in memory.

pub fn with_replacement_strategy(self, strategy: Replacement) -> Self

What rules to use when choosing whether to overwrite the current value in the transposition table.

pub fn with_null_window_search(self, null: bool) -> Self

Whether to add null-window searches to try to prune branches that are probably worse than those already found. Also known as principal variation search.

pub fn with_null_move_depth(self, depth_reduction: u8) -> Self

Whether to attempt to cut off early by seeing if each node is amazing even after passing the turn to the opponent. Null move search explores the tree at a depth reduced by this amount.

pub fn with_singular_extension(self) -> Self

Whether to extend a branch of the search (by 1) if there is only one move (or only one reasonable move).

pub fn with_aspiration_window(self, window: Evaluation) -> Self

Whether to search first in a narrow window around the previous root value on each iteration.

pub fn with_mtdf(self) -> Self

Whether to search for the correct value in each iteration using only null-window “Tests”, with the MTD(f) algorithm. Can be more efficient if the evaluation function is coarse grained.

pub fn with_double_step_increment(self) -> Self

Increment the depth by two between iterations.

pub fn with_quiescence_search_depth(self, depth: u8) -> Self

Enable quiescence search at the leaves of the search tree. The Evaluator must implement generate_noisy_moves for the search to know when the state has become “quiet”.

pub fn with_min_reorder_moves_depth(self, depth: u8) -> Self

Enable the Evaluator’s move reordering after generating moves for all nodes at this depth or higher. Reordering can be an expensive operation, but it could cut off a lot of nodes if done well high in the search tree.

pub fn with_countermoves(self) -> Self

Enable the countermove table, which reorders to the front moves that have worked to counter the previous move in other branches.

pub fn with_countermove_history(self) -> Self

Enable the countermove history table. It keeps a counter for moves that have caused beta cutoffs in other branches, and reorders moves based on this counter.

pub fn verbose(self) -> Self

Enable verbose print statements of the ongoing performance of the search.

Trait Implementations

impl Clone for IterativeOptions

fn clone(&self) -> IterativeOptions

Returns a copy of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Default for IterativeOptions

fn default() -> Self

Returns the “default value” for a type.

impl Copy for IterativeOptions