Applying the displacement resolver has no effect when both variables are non-singleton.
Suppose there are two variables v0∈{a,b,c} and v1∈{a,b}, enforcing the constraint v0≠v1 has no effect on the domains of the v0 and v1 variables hence, once applied the resolver, the flaw will reemerge at the next iteration.
In case of non-singleton variables, the displacement resolver could be replaced with n placement resolvers. In the above case, for example, the displacement resolver would be replaced with those placement resolvers which would assign different values to the v0 and v1 variables, i.e. <a,b>, <b,a>, <c,a> and <c,b>.
Placement resolvers are heavy committing resolvers and should hence be despised, compared to other resolvers..
Applying the displacement resolver has no effect when both variables are non-singleton.
Suppose there are two variables
v0∈{a,b,c}
andv1∈{a,b}
, enforcing the constraintv0≠v1
has no effect on the domains of thev0
andv1
variables hence, once applied the resolver, the flaw will reemerge at the next iteration.In case of non-singleton variables, the displacement resolver could be replaced with
n
placement resolvers. In the above case, for example, the displacement resolver would be replaced with those placement resolvers which would assign different values to thev0
andv1
variables, i.e.<a,b>
,<b,a>
,<c,a>
and<c,b>
.Placement resolvers are heavy committing resolvers and should hence be despised, compared to other resolvers..