view release on metacpan or  search on metacpan

lib/AI/Pathfinding/OptimizeMultiple.pm  view on Meta::CPAN

            ( $ones_constant x $flares_num_iters ) + (
                PDL::MatrixOps::identity( $self->_get_num_scans() ) *
                    $iters_quota
            )
        );

        # print "\$next_num_iters = $next_num_iters\n";

        my $iters = $self->_scans_data()->slice(":,:,0");

        my $iters_repeat =
            $iters->dummy( 0, $self->_get_num_scans() )->xchg( 1, 2 )
            ->clump( 2 .. 3 );

        # print "\$iters_repeat =", join(",",$iters_repeat->dims()), "\n";

        my $next_num_iters_repeat =
            $next_num_iters->dummy( 0, $self->_num_boards() )->xchg( 0, 2 );

# print "\$next_num_iters_repeat =", join(",",$next_num_iters_repeat->dims()), "\n";

        # A boolean tensor of which boards were solved:
        # Dimension 0 - Which scan is it. - size - _get_num_scans()
        # Dimension 1 - Which scan we added the quota to
        #   - size - _get_num_scans()
        # Dimension 2 - Which board. - size - _num_boards()
        my $solved =
            ( $iters_repeat >= 0 ) * ( $iters_repeat < $next_num_iters_repeat );

      # print "\$num_moves_repeat =", join(",",$num_moves_repeat->dims()), "\n";

        my $num_moves_solved =
            ( $solved * $num_moves_repeat ) +
            ( $solved->not() * $UNSOLVED_NUM_MOVES_CONSTANT );

        my $minimal_num_moves_solved =
            $num_moves_solved->xchg( 0, 1 )->minimum();

        my $which_minima_are_solved =
            ( $minimal_num_moves_solved != $UNSOLVED_NUM_MOVES_CONSTANT );

        my $minimal_with_zeroes =
            $which_minima_are_solved * $minimal_num_moves_solved;

        my $solved_moves_sums   = _my_xchg_sum_over($minimal_with_zeroes);
        my $solved_moves_counts = _my_xchg_sum_over($which_minima_are_solved);
        my $solved_moves_avgs   = $solved_moves_sums / $solved_moves_counts;

        # print join(",", $solved_moves_avgs->minmaximum()), "\n";

        my $min_avg;

        ( $min_avg, undef, $selected_scan_idx, undef ) =
            $solved_moves_avgs->minmaximum();

        $last_avg = $min_avg;

        push @{ $self->chosen_scans() },
            $self->_calc_chosen_scan( $selected_scan_idx, $iters_quota );

        $flares_num_iters->set( $selected_scan_idx,
            $flares_num_iters->at($selected_scan_idx) + $iters_quota );
        $self->_selected_scans()->[$selected_scan_idx]->mark_as_used();

        $iters_quota = 0;

        my $num_solved = $solved_moves_counts->at($selected_scan_idx);

        my $flares_num_iters_repeat =
            $flares_num_iters->dummy( 0, $self->_num_boards() );

        # A boolean tensor:
        # Dimension 0 - board.
        # Dimension 1 - scans.
        my $solved_with_which_iter =
            ( $flares_num_iters_repeat >= $iters->clump( 1 .. 2 ) ) &
            ( $iters->clump( 1 .. 2 ) >= 0 );

        my $total_num_iters = (
            ( $solved_with_which_iter * $flares_num_iters_repeat )->sum() + (
                $solved_with_which_iter->not()->andover() *
                    $flares_num_iters->sum()
            )->sum()
        );

        print "Finished ", $loop_iter_num++,
" ; #Solved = $num_solved ; Iters = $total_num_iters ; Avg = $min_avg\n";
        STDOUT->flush();
    }
}

sub calc_board_iters
{
    my $self  = shift;
    my $board = shift;

    my $board_iters = 0;

    my @info      = PDL::list( $self->_orig_scans_data()->slice("$board,:") );
    my @orig_info = @info;

    foreach my $s ( @{ $self->chosen_scans() } )
    {
        if (   ( $info[ $s->scan_idx() ] > 0 )
            && ( $info[ $s->scan_idx() ] <= $s->iters() ) )
        {
            $board_iters += $info[ $s->iters() ];
            last;
        }
        else
        {
            if ( $info[ $s->scan_idx() ] > 0 )
            {
                $info[ $s->scan_idx() ] -= $s->iters();
            }
            $board_iters += $s->iters();
        }
    }

    return {
        'per_scan_iters' => \@orig_info,

lib/AI/Pathfinding/OptimizeMultiple.pm  view on Meta::CPAN

        {
            status      => $status,
            scan_runs   => \@scan_runs,
            total_iters => $board_iters,
        }
    );
}

sub _trace
{
    my ( $self, $args ) = @_;

    if ( my $trace_callback = $self->_trace_cb() )
    {
        $trace_callback->($args);
    }

    return;
}

sub get_total_iters
{
    my $self = shift;

    return $self->_total_iters();
}

sub _add_to_total_iters
{
    my $self = shift;

    my $how_much = shift;

    $self->_total_iters( $self->_total_iters() + $how_much );

    return;
}

sub _add_to_total_boards_solved
{
    my $self = shift;

    my $how_much = shift;

    $self->_total_boards_solved( $self->_total_boards_solved() + $how_much );

    return;
}

1;    # End of AI::Pathfinding::OptimizeMultiple

__END__

=pod

=encoding UTF-8

=head1 NAME

AI::Pathfinding::OptimizeMultiple - optimize path finding searches for a large
set of initial conditions (for better average performance).

=head1 VERSION

version 0.0.17

=head1 SYNOPSIS

    use AI::Pathfinding::OptimizeMultiple

    my @scans =
    (
        {
            name => "first_search"
        },
        {
            name => "second_search",
        },
        {
            name => "third_search",
        },
    );

    my $obj = AI::Pathfinding::OptimizeMultiple->new(
        {
            scans => \@scans,
            num_boards => 32_000,
            optimize_for => 'speed',
            scans_iters_pdls =>
            {
                first_search => $first_search_pdl,
                second_search => $second_search_pdl,
            },
            quotas => [400, 300, 200],
            selected_scans =>
            [
                AI::Pathfinding::OptimizeMultiple::Scan->new(
                    id => 'first_search',
                    cmd_line => "--preset first_search",
                ),
                AI::Pathfinding::OptimizeMultiple::Scan->new(
                    id => 'second_search',
                    cmd_line => "--preset second_search",
                ),
                AI::Pathfinding::OptimizeMultiple::Scan->new(
                    id => 'third_search',
                    cmd_line => "--preset third_search",
                ),
            ],
        }
    );

    $obj->calc_meta_scan();

    foreach my $scan_alloc (@{$self->chosen_scans()})
    {
        printf "Run %s for %d iterations.\n",
            $scans[$scan_alloc->scan_idx], $scan_alloc->iters;
    }

=head1 DESCRIPTION

This CPAN distribution implements the algorithm described here:

=over 4

=item * L<https://groups.google.com/group/comp.ai.games/msg/41e899e9beea5583?dmode=source&output=gplain&noredirect>

=item * L<http://www.shlomifish.org/lecture/Perl/Lightning/Opt-Multi-Task-in-PDL/>

=back

Given statistics on the performance of several game AI searches (or scans)
across a representative number of initial cases, find a scan
that solves most deals with close-to-optimal performance, by using switch
tasking.

=head1 SUBROUTINES/METHODS

=head2 my $chosen_scans_array_ref = $self->chosen_scans()

Returns the scans that have been chosen to perform the iteration. Each one is
a AI::Pathfinding::OptimizeMultiple::ScanRun object.

=head2 $calc_meta_scan->calc_meta_scan()

Calculates the meta-scan after initialisation. See here for the details
of the algorithm:

L<http://www.shlomifish.org/lecture/Freecell-Solver/The-Next-Pres/slides/multi-tasking/best-meta-scan/>

=head2 $self->calc_flares_meta_scan()

This function calculates the flares meta-scan: i.e: assuming that all atomic
scans are run one after the other and the shortest solutions of all
successful scans are being picked.

=head2 $calc_meta_scan->calc_board_iters($board_idx)

Calculates the iterations of the board $board_idx in all the scans.

Returns a hash_ref containing the key 'per_scan_iters' for the iterations
per scan, and 'board_iters' for the total board iterations when ran in the
scans.

=head2 my $status = $calc_meta_scan->get_final_status()

Returns the status as string:

=over 4

=item * "solved_all"

=item * "iterating"

=item * "out_of_quotas"

=back

=head2 my $sim_results_obj = $calc_meta_scan->simulate_board($board_idx, $args)

Simulates the board No $board_idx through the scan. Returns a
L<AI::Pathfinding::OptimizeMultiple::SimulationResults> object.

$args is an optional hash reference. It may contain a value with the key of
C<'chosen_scans'> that may specify an alternative scans to traverse.

=head2 my $n = $calc_meta_scan->get_total_iters()

Returns the total iterations count so far.

=head2 BUILD()

Moo leftover. B<INTERNAL USE>.

=head1 SEE ALSO

=over 4

=item * L<Freecell Solver|http://fc-solve.shlomifish.org/>

For which this code was first written and used.

=item * L<Alternative Implementation in C#/.NET|https://bitbucket.org/shlomif/fc-solve/src/cc5b428ed9bad0132d7a7bc1a14fc6d3650edf45/fc-solve/presets/soft-threads/meta-moves/auto-gen/optimize-seq?at=master>

An Alternative implementation in C#/.NET, which was written because the
performance of the Perl/PDL code was too slow.

=item * L<PDL> - Perl Data Language

Used here.

=back

=head1 AUTHOR

Shlomi Fish, L<http://www.shlomifish.org/> .

=head1 ACKNOWLEDGEMENTS

B<popl> from Freenode's #perl for trying to dig some references to an existing
algorithm in the scientific literature.

=for :stopwords cpan testmatrix url bugtracker rt cpants kwalitee diff irc mailto metadata placeholders metacpan

=head1 SUPPORT

=head2 Websites

The following websites have more information about this module, and may be of help to you. As always,
in addition to those websites please use your favorite search engine to discover more resources.

=over 4

=item *

MetaCPAN

A modern, open-source CPAN search engine, useful to view POD in HTML format.

L<https://metacpan.org/release/AI-Pathfinding-OptimizeMultiple>

=item *

RT: CPAN's Bug Tracker

The RT ( Request Tracker ) website is the default bug/issue tracking system for CPAN.

L<https://rt.cpan.org/Public/Dist/Display.html?Name=AI-Pathfinding-OptimizeMultiple>

=item *

CPANTS

The CPANTS is a website that analyzes the Kwalitee ( code metrics ) of a distribution.

L<http://cpants.cpanauthors.org/dist/AI-Pathfinding-OptimizeMultiple>

=item *

CPAN Testers

The CPAN Testers is a network of smoke testers who run automated tests on uploaded CPAN distributions.