AIMMS to ODH Mapping

ODH engine options

The table below shows in the left column the AIMMS options that control the ODH engine; the right column displays for each AIMMS options the associated ODH-CPLEX parameter.

Option name in AIMMS	Name in ODH-CPLEX
Backtrack Limit	MAXBACKTRACK
Decomposition Density	DECOMPDENSITY
Global Bounds	GLOBALBOUNDS
Initial Divisor Value	INTERDIV
Initial Divisor Value Sub Models	DIVISOR
Maximum Divisor Repeats	MAXREPEAT
Maximum Divisor Repeats Infeasible	MAXINFREPEAT
Maximum Divisor Value	MAXINTERDIV
Optimization Method	FEASOPT
Write Decomposition File
Clean Variables Sub Models	VARIABLECLEAN
Objective Target	OBJTARGET
ODH Feasibility Tolerance	FEASTOL
ODH Presolve	ODHPRESOLVE
ODH Seed	SEED
Presolve	PRESOLVE
Quick First Solve	QUICKFIRSTSOLVE
Reject Infeasible Solutions	REJECTINFSOL
Relax SOS2	RELAXSOS2
Remove Infeasibilities Method	PHASE12
Search Mode
Write Solution File	WRITESOLUTION
First Feasible Heuristic	FIRSTFEAS
First Feasible Heuristic Continue	FIRSTFEASCONTINUE
First Feasible Heuristic Effort Level	FIRSTFEASEFFORT
First Feasible Heuristic Shift	FIRSTFEASSHIFT
Recurse	RECURSE
Recurse Decomposition Method	RECURSEDECOMP
Recurse Iteration Limit	RECURSEITERLIM
Recurse Iteration Limit Solution	RECURSESOLITERLIM
Recurse Minimum Iterations	RECURSEMINITERLIM
Solution Improvement Heuristic Mode	DETERMINISTIC
Solution Improvement Heuristic Penalty	PENALTY
Solution Improvement Heuristic Strategy	STRATEGY
Recurse Log	RECURSELOG
Status Display
Thread Log	THREADLOG
Thread Limit	THREADS
Thread Synchronization Frequency	SYNCFREQ
Processor Lock	PROCESSORLOCK

CPLEX solve options

The two tables below show options used by ODH-CPLEX for the main CPLEX solve in ODH-CPLEX (if the Search Mode is set to ‘Global Solution’) or for getting an initial feasible solution (if the Search Mode is set to ‘Local Solution’)

Option name in AIMMS	Name in ODH-CPLEX for main CPLEX solve
Barrier Algorithm	CPX_BARALG
Barrier Convergence Tolerance	CPX_BAREPCOMP
Barrier Crossover Algorithm	CPX_BARCROSSALG
Barrier Density Definition	CPX_BARCOLNZ
Barrier Growth Limit	CPX_BARGROWTH
Barrier Iterations	CPX_BARITLIM
Barrier Maximal Number of Corrections	CPX_BARMAXCOR
Barrier Objective Range	CPX_BAROBJRNG
Barrier Ordering	CPX_BARORDER
Barrier Start Algorithm	CPX_BARSTARTALG
Advanced Start	CPX_ADVIND
Cleanup Coefficients
Clock Type	CPX_CLOCKTYPE
Conflict Algorithm	CPX_CONFLICTALG
Data Check and Modeling Assistance	CPX_DATACHECK
Deterministic Time Limit	CPX_DETTILIM
LP File
Memory Emphasis	CPX_MEMORYEMPHASIS
MPS
Numerical Emphasis	CPX_NUMERICALEMPHASIS
Ord File
Random Seed	CPX_RANDOMSEED
Round Coefficients
Sav File
Scale Sifting Algorithm	CPX_SCAINDCPX_SIFTALG
Updates Batch Size
Barrier Display	CPX_BARDISPLAY
Clone Log Files	CPX_CLONELOG
MIP Display	CPX_MIPDISPLAY
MIP Interval	CPX_MIPINTERVAL
Simplex Display	CPX_SIMDISPLAY
Backtrack	CPX_BTTOL
Branch	CPX_BRDIR
Difference Objective	CPX_OBJDIF
Integrality	CPX_EPINT
Maximal Number of Nodes	CPX_NODELIM
MIP Basis
MIP Candidate List	CPX_STRONGCANDLIM
MIP Dive Strategy	CPX_DIVETYPE
MIP Emphasis	CPX_MIPEMPHASIS
MIP Method	CPX_SUBALG
MIP Probing	CPX_PROBE
MIP Search Strategy	CPX_MIPSEARCH
MIP Start Algorithm	CPX_STARTALG
MIP Tree Memory Limit	CPX_TRELIM
MIP Update
Node File	CPX_NODEFILEIND
MIP Node Presolve	CPX_PRESLVND
Number of Parallel Threads	CPX_STRONGTHREADLIM
Number of Repair Attempts	CPX_REPAIRTRIES
Number of Simplex Iterations MIP Priority Order Switch MIP Priority Order Type	CPX_STRONGITLIMCPX_MIPORDINDCPX_MIPORDTYPE
Probing Time	CPX_PROBETIME
Probing Time Deterministic	CPX_PROBEDETTIME
Relative Difference Objective	CPX_RELOBJDIF
Select Variables	CPX_VARSEL
Selection of Nodes	CPX_NODESEL
Working Memory Limit	CPX_WORKMEM
Write MIP Starts
SubMIP Node Limit	CPX_SUBMIPNODELIMIT
SubMIP Scale	CPX_SUBMIPSCAIND
SubMIP Start Algorithm	CPX_SUBMIPSTARTALG
SubMIP Subproblem Algorithm	CPX_SUBMIPSUBALG
BQP Cuts	CPX_BQPCUTS
Clique Cuts	CPX_CLIQUES
Cover Cuts	CPX_COVERS
Cut Generation Limit	CPX_AGGCUTLIM
Cut Limit	CPX_EACHCUTLIM
Cuts Factor	CPX_CUTSFACTOR
Disjunctive Cuts	CPX_DISJCUTS
Flow Cover Cuts	CPX_FLOWCOVERS
Flow Path Cuts	CPX_FLOWPATHS
Gomory Cuts	CPX_FRACCUTS
Gomory Cuts Candidate Limit	CPX_FRACCAND
Gomory Cuts Pass Limit	CPX_FRACPASS
GUB Cover Cuts	CPX_GUBCOVERS
Implied Bound Cuts	CPX_IMPLBD
Lift and Project Cuts	CPX_LANDPCUTS
Local Implied Bound Cuts	CPX_LOCALIMPLBD
MCF Cuts	CPX_MCFCUTS
MIP Number of Cut Passes	CPX_CUTPASS
Mixed Integer Rounding Cuts	CPX_MIRCUTS
Node Cuts	CPX_NODECUTS
RLT Cuts	CPX_RLTCUTS
Feasibility Pump Heuristic	CPX_FPHEUR
Heuristic Effort	CPX_HEUREFFORT
Heuristic Frequency	CPX_HEURFREQ
Local Branching Heuristic	CPX_LBHEUR
RINS Heuristic Frequency	CPX_RINSHEUR
Zero Half Cuts	CPX_ZEROHALFCUTS
Boundstrength	CPX_BNDSTRENIND
Coefficient Reduction	CPX_COEREDIND
Preprocessing Symmetry	CPX_SYMMETRY
Presolve Relaxed MIP	CPX_RELAXPREIND
Repeat Presolve	CPX_REPEATPRESOLVE
SOS1 Reformulations	CPX_SOS1REFORM
SOS2 Reformulations	CPX_SOS2REFORM
Polishing Absolute MIP Gap	CPX_POLISHAFTEREPAGAP
Polishing Number of Nodes	CPX_POLISHAFTERNODE
Polishing Number of Solutions	CPX_POLISHAFTERINTSOL
Polishing Relative MIP Gap	CPX_POLISHAFTEREPGAP
Polishing Time	CPX_POLISHAFTERTIME
Polishing Time Deterministic	CPX_POLISHAFTERDETTIME
Do Populate
Pool Absolute Objective Gap	CPX_SOLNPOOLAGAP
Pool Capacity	CPX_SOLNPOOLCAPACITY
Pool Intensity	CPX_SOLNPOOLINTENSITY
Pool Relative Objective Gap	CPX_SOLNPOOLGAP
Pool Replacement Strategy	CPX_SOLNPOOLREPLACE
Populate Time Limit
Population Limit	CPX_POPULATELIM
Network Feasibility	CPX_NETEPRHS
Network Iterations	CPX_NETITLIM
Network Optimality	CPX_NETEPOPT
Network Pricing	CPX_NETPPRIIND
Auxiliary Root Threads	CPX_AUXROOTTHREADS
Global Thread Limit	CPX_THREADS
Parallel Mode	CPX_PARALLELMODE
Aggregator	CPX_AGGIND
Dependency Folding	CPX_DEPINDCPX_FOLDING
Limit Substitutions	CPX_AGGFILL
Number of Iterations in Presolve	CPX_PREPASS
Preprocessing Reduction Types	CPX_REDUCE
Presolve Pass Dual	CPX_PREDUAL
Adjust MIQP	CPX_QPMAKEPSDIND
Barrier Convergence Tolerance for QCP	CPX_BARQCPEPCOMP
MIQCP Strategy	CPX_MIQCPSTRAT
QP Linearization	CPX_QTOLININD
QP Method	CPX_QPMETHOD
QP Nonzeros Read Limit	CPX_QPNZREADLIM
Solution Target	CPX_OPTIMALITYTARGET
Crash Ordering	CPX_CRAIND
Dual Pricing Algorithm	CPX_DPRIIND
Dynamic Row Management	CPX_DYNAMICROWS
Feasibility	CPX_EPRHS
Markowitz	CPX_EPMRK
Optimality	CPX_EPOPT
Perturbation Constant	CPX_EPPER
Perturbation Indicator	CPX_PERIND
Pricing	CPX_PRICELIM
Primal Pricing Algorithm	CPX_PPRIIND
Refactor	CPX_REINV
Sifting from Simplex	CPX_SIFTSIM
Singular	CPX_SINGLIM
Stalled Iterations	CPX_PERLIM

The table below shows Solvers General options that are mapped to ODH-CPLEX parameters.

Option name in AIMMS	Name in ODH-CPLEX
Cutoff	CPX_CUTLO
	CPX_CUTUP
Maximal Number of Integer Solutions	CPX_INTSOLLIM
MIP Absolute Optimality Tolerance	CPX_EPAGAP
MIP Relative Optimality Tolerance	CPX_EPGAP
Iteration Limit	CPX_ITLIM
Time Limit	TIMELIMIT

Heuristic sub-model options

The heuristic sub-model parameters can only be set using a Parameter File . The syntax for the parameters that influence the heuristic sub-model CPLEX solves is the following: SUB_<parameter> where <parameter> refers to the ODH-CPLEX name in the second table above. The syntax for the parameters that influence the heuristic sub-model CPLEX solves in Phase I is the following: PHASE1_<parameter>.

For example, SUB_CPX_RINSHEUR specifies the RINS heuristic frequency for CPLEX if it used to solve a heuristic sub-model, while PHASE1_CPX_RINSHEUR specifies the RINS heuristic frequency for CPLEX if it used to solve a heuristic sub-model in Phase I.

Learn more about

• Parameter File

• Search Mode