Repository navigation
Network simplex options missing in CPLEX MP #251
Description
ASL's "netopt" is described in the option listing as
alg:network (network, netopt)
Solve (substructure of) (MIP node) LPs by network simplex method.
whereas it should correspond to the listing for the ASL version:
netopt 0 means never invoke the network optimizer.
1 (default) means invoke the network optimizer
only if the model had node and arc declarations.
2 means always invoke the network optimizer
(unless there are integer variables); the network
optimizer may be able to find and exploit an
embedded network.
3 is similar to 2, but sets CPLEX's LPMethod
to CPX_ALG_NET rather than explicitly invoking
the network optimizer. This might make a
difference if CPLEX's presolve makes relevant
reductions.
ASL's other network simplex options -- all the other ones that begin with "net" -- are missing from the MP implementation.
(This was reported by a user who upgraded CPLEX and could no longer use netopt=2 and netfind=2 except by switching to cplexasl.)
Activity
glebbelov commentedon Feb 11, 2025
MP does not read node/arc declarations from the NL model currently. Thus, the network structure is not known and the network optimizer cannot be called in general; only CPX_ALG_NET can be used with the usual LP solver. To achieve best network simplex performance,
presolve=0in cplex might be needed.MP's
netopt(no value) equals ASL'snetopt=3. @4er4er4er do we need ASL-compatible values, e.g., fornetoptonly? @mapgccv Why hasnetoptno values, along withalg:primal,alg:barrier, etc? (They areprimalopt,baroptin ASL by the way.)netfindshould be easy to add.4er4er4er commentedon Feb 11, 2025
It seems to be more complicated than I thought. When there are node and arc definitions, the output from CPLEX (with default settings) suggests that it received a network structure from ASL. But when there are no node and arc definitions -- as in multi.mod + multi.dat from the book -- netopt 2 activates the network extractor in CPLEX, which looks for a network structure as specified by netfind:
It would be interesting to make a version of this model using node and arc definitions for Supply and Demand, plus subject-to definitions for the Multi constraints, and see what happens with various netopt and netfind settings.