Please could people advise me on whether I need to start the program again.

Ditto as above wrt what all the others wrote. Your program makes my eyes bleed, period. Thus if you ask me, then yes: your should rewrite it from scratch. Of course you should first learn how to do so rather than posting either the same very snippets over and over again or whole programs with vague rfcs attached to them. Consider:

#!/usr/bin/perl

use strict;
use warnings;
use Data::Dumper 'Dumper';

my $dir = "csv/";

my @filenames = grep /csv/ && -f "$dir/$_" => do {
    opendir my $dh, $dir or die "Can't opendir `$dir': $!\n";
    readdir $dh;
};

my @place;

open my $prep, "$dir/$filenames[0]" or
  die "Can't open file `$filenames[0]': $!\n";

while (<$prep>) {
    next unless /^([^,]+),([^,]+),(?:\d+,){4},(?:\d+,){4},(?:\d+,){3}/
+;
    my $place = "$1_$2";
    my $reglookup = "$1,$2";
    print "\n", '>' x 21, "\n", $reglookup, "\n", '>' x 21, "\n";
    y/ /_/ => y/\"//d for $place_tag_string;
    s/ /\\s/g => s/([,\"()&])/\\$1/ge for $reglookup_tag_string;
    push @place, 
      { type => $1, name => $2, tag => $place, reg => $reglookup };
}
close $prep;

my $filenumber;
for (@place) {
    print "\n@@@@@@@@@@\n\n" $_->{reg}, "\n\n";
    for my $file (@file_names) {
        my ($rl, $pt) = @{ $_ }{qw/reg tag/};
        $filenumber++;
        print "\n\n\n\n$rl, $dir, $file, $filenumber, $pt\n\n\n\n";
        print_out_files($rl, $dir, $filename, $filenumber, $pt);
    }
}

sub print_out_files {
    my %files;
    my ($rl, $dir, $file, $filenumber, $pt) = @_;
    my $Is_it_printed = 0;

    $files{Outfile}   = "Result/Result_$pt.txt";
    $files{Outfile_B} = "Result/Result_B_$pt.txt";

    open my $ofh, '+>', $files{Outfile} or
      die "Can't create output file `$files{Outfile}': $!\n";
    open my $ofh, '+>', $files{Outfile_B} or
      die "Can't create output file `$files{Outfile_B}': $!\n";
[download]

Which I got out GrandFather's improved version of your program by further:

• replacing the shebang line with one that makes sense;
• removing unnecessary stuff;
• declaring variables as close as possible to where they're used;
• using saner opens and opendirs;
• using low precedency logical operators for flow control, as is almost always best done;
• replacing the awkward /^.*.csv/ with the equivalent /csv/, provided that that is what you want;
• using a single @place array instead of the three ones you were;
• using more reasonable variable names that are neither extremely short nor too long and error-prone to remember and hard to parse when reading;
• removing the unnecessary and confusing $_ =~;
• writing a hopefully saner regex of the huge monster, (but then is a single regex the best way to do it? I don't think it is and I would go the split way instead, but that's too much of an intervention);
• using next for flow control rather than a huge if block;
• using tr instead of s where appropriate;
• using aliasing for's to avoid =~ madness;
• removing an unnecessary next unless defined $_ in a place where $_ can't be undefined - similarly later;
• removing unnecessary operations, like yet another s/ /\\s/g on something which had already undergone it;
• removing clumsy $array_position logic and also putting @reglookup data into @place
- d'oh, you made me step back some lines;

This is not a complete rewrite because arrived at the point at which I stopped I had grown completely tired, the rest of your code scared me and hey, nobody is paying me for this... I hope you get the picture and can go on yourself.

Consider that where I stopped I got 53 lines of reasonably terse code out of your original 99 ones. And I didn't want to change the logic too much, too.