How big are your arrays? An CPU inefficient but memory efficient version is:

my @a_indices = (1,3,5);
my @b = map {my $t = $_; (grep {$t == $_} @a_indices) ? undef : $t} 0 
+.. 9;

print join ', ', grep {defined} @b;
[download]

Prints:

0, 2, 4, 6, 7, 8, 9
[download]

On the other hand:

my @a_indices = (1,3,5);
my %hasIndex = map {$_ => undef} @a_indices;
my @b = map {exists $hasIndex{$_} ? undef : $_} 0 .. 9;
[download]

is likely to be faster for large arrays, but at the expense of creating a hash.