NLP for CP
Addressing Constraint Programming with Natural Language Processing
Home
Resources
Publications
Correct
predictions are in
blue
. If we detect only a subset of a labelled sentence, we highlight the caught part as
blue
, the missing part
light blue.
False positives
are in
green
and
false negatives
are in
red
.
Problem 108 (Winner Determination Problem) — Constraint detection
A
set
of
bidders
is
denoted
by
N
=
-LCB-
1
,
...
,
n
-RCB-
,
a
set
of
items
by
M
=
-LCB-
1
,
...
,
m
-RCB-
.
A
bundle
S
is
a
set
of
items
:
$
S
\
subset
M$
.
For
a
bundle
S
and
a
bidder
i
,
we
denote
by
vi
-LRB-
S
-RRB-
the
package
bid
that
bidder
i
makes
for
bundle
S
,
i.e.
,
the
maximal
price
that
i
announces
to
be
willing
to
pay
for
S
.
An
allocation
of
the
items
is
described
by
variables
$
xi
-LRB-
S
-RRB-
\
in
-LCB-
0
,
1
-RCB-
$
.
The
variable
xi
-LRB-
S
-RRB-
is
equal
to
one
if
and
only
if
bidder
i
gets
bundle
S
.
An
allocation
$
-LRB-
xi
-LRB-
S
-RRB-
|
i
\
in
N
,
S
\
subset
M
-RRB-
$
is
said
to
be
feasible
if
it
allocates
no
item
more
than
once
:
$
\
sum
_
-LCB-
i
\
in
N
-RCB-
\
sum
_
-LCB-
S
\
subset
M
,
j
\
in
S
-RCB-
x_i
-LRB-
S
-RRB-
\
le
1
\
forall
j
\
in
M$
,
and
at
most
one
subset
to
every
bidder
:
$
\
sum
_
-LCB-
S
\
subset
M
-RCB-
x_i
-LRB-
S
-RRB-
\
le
1
\
forall
i
\
in
N$
.
Given
bids
$
v_i
,
i
=
1
,
...
,
n
$
,
the
winner
determination
problem
is
the
problem
to
compute
$
x
\
in
argmax
-LRB-
\
sum
_
-LCB-
i
\
in
N
-RCB-
v_i
-LRB-
S
-RRB-
x_i
-LRB-
S
-RRB-
|
x
is
a
feasible
allocation
-RRB-
$
.
Problem 108 (Winner Determination Problem) — Detection of the decisions and objects to be modeled
A
set
of
bidders
is
denoted
by
N
=
-LCB-
1
,
...
,
n
-RCB-
,
a
set
of
items
by
M
=
-LCB-
1
,
...
,
m
-RCB-
.
A
bundle
S
is
a
set
of
items
:
$
S
\
subset
M$
.
For
a
bundle
S
and
a
bidder
i
,
we
denote
by
vi
-LRB-
S
-RRB-
the
package
bid
that
bidder
i
makes
for
bundle
S
,
i.e.
,
the
maximal
price
that
i
announces
to
be
willing
to
pay
for
S
.
An
allocation
of
the
items
is
described
by
variables
$
xi
-LRB-
S
-RRB-
\
in
-LCB-
0
,
1
-RCB-
$
.
The
variable
xi
-LRB-
S
-RRB-
is
equal
to
one
if
and
only
if
bidder
i
gets
bundle
S
.
An
allocation
$
-LRB-
xi
-LRB-
S
-RRB-
|
i
\
in
N
,
S
\
subset
M
-RRB-
$
is
said
to
be
feasible
if
it
allocates
no
item
more
than
once
:
$
\
sum
_
-LCB-
i
\
in
N
-RCB-
\
sum
_
-LCB-
S
\
subset
M
,
j
\
in
S
-RCB-
x_i
-LRB-
S
-RRB-
\
le
1
\
forall
j
\
in
M$
,
and
at
most
one
subset
to
every
bidder
:
$
\
sum
_
-LCB-
S
\
subset
M
-RCB-
x_i
-LRB-
S
-RRB-
\
le
1
\
forall
i
\
in
N$
.
Given
bids
$
v_i
,
i
=
1
,
...
,
n
$
,
the
winner
determination
problem
is
the
problem
to
compute
$
x
\
in
argmax
-LRB-
\
sum
_
-LCB-
i
\
in
N
-RCB-
v_i
-LRB-
S
-RRB-
x_i
-LRB-
S
-RRB-
|
x
is
a
feasible
allocation
-RRB-
$
.
Back to list