<s>
In	O
mathematical	O
optimization	O
,	O
neighborhood	B-Algorithm
search	I-Algorithm
is	O
a	O
technique	O
that	O
tries	O
to	O
find	O
good	O
or	O
near-optimal	O
solutions	O
to	O
a	O
combinatorial	O
optimisation	O
problem	O
by	O
repeatedly	O
transforming	O
a	O
current	O
solution	O
into	O
a	O
different	O
solution	O
in	O
the	O
neighborhood	O
of	O
the	O
current	O
solution	O
.	O
</s>
<s>
For	O
a	O
very	B-Algorithm
large-scale	I-Algorithm
neighborhood	I-Algorithm
search	I-Algorithm
,	O
the	O
neighborhood	O
is	O
large	O
and	O
possibly	O
exponentially	O
sized	O
.	O
</s>
<s>
If	O
neighborhood	O
searched	O
is	O
limited	O
to	O
just	O
one	O
or	O
a	O
very	O
small	O
number	O
of	O
changes	O
from	O
the	O
current	O
solution	O
,	O
then	O
it	O
can	O
be	O
difficult	O
to	O
escape	O
from	O
local	O
minima	O
,	O
even	O
with	O
additional	O
meta-heuristic	O
techniques	O
such	O
as	O
Simulated	B-Algorithm
Annealing	I-Algorithm
or	O
Tabu	B-Algorithm
search	I-Algorithm
.	O
</s>
<s>
In	O
large	O
neighborhood	B-Algorithm
search	I-Algorithm
techniques	O
,	O
the	O
possible	O
changes	O
from	O
one	O
solution	O
to	O
its	O
neighbor	O
may	O
allow	O
tens	O
or	O
hundreds	O
of	O
values	O
to	O
change	O
,	O
and	O
this	O
means	O
that	O
the	O
size	O
of	O
the	O
neighborhood	O
may	O
itself	O
be	O
sufficient	O
to	O
allow	O
the	O
search	O
process	O
to	O
avoid	O
or	O
escape	O
local	O
minima	O
,	O
though	O
additional	O
meta-heuristic	O
techniques	O
can	O
still	O
improve	O
performance	O
.	O
</s>
