[算法知识]Gurobi建模求解相关参数设置

Parameters fall into the following categories:

• Termination: affect the termination of an optimize() call

• BarIterLimit: limits the number of barrier iterations performed

• BestBdStop: sets a best bound values at which optimization should stop

• BestObjStop: sets an objective value at which optimization should stop

• Cutoff: sets a target objective value

• IterationLimit: limits the number of simplex iterations performed

• NodeLimit: limits the number of MIP nodes explored

• SolutionLimit: sets a target for the number of feasible solutions found

• TimeLimit: limits the total time expended (in seconds)求解时间限制

• Tolerances: control the allowable feasibility or optimality violations

• BarConvTol: barrier convergence tolerance
  vBarQCPConvTol: barrier convergence tolerance for QCP models

• FeasibilityTol: primal feasibility tolerance

• IntFeasTol: integer feasibility tolerance

• MarkowitzTol: threshold pivoting tolerance

• MIPGap: target relative MIP optimality gap

• MIPGapAbs: target absolute MIP optimality gap
  求解Gap限制，默认为10e-4

• OptimalityTol: dual feasibility tolerance

• PSDTol: QP positive semidefinite tolerance

• Simplex: affect the simplex algorithms

• InfUnbdInfo: makes additional information available for infeasible or unbounded LP models

• NormAdjust: chooses different pricing norm variants

• ObjScale: controls objective scaling

• PerturbValue: controls the magnitude of any simplex perturbations

• Quad: turns quad precision on or off

• ScaleFlag: turns model scaling on or off

• Sifting: dual simplex sifting strategy for LP, MIP root and MIP nodes

• SiftMethod: chooses from dual, primal and barrier to solve sifting subproblems

• SimplexPricing: determines variable pricing strategy

• Barrier: affect the barrier algorithms

• BarCorrectors: limits the number of central corrections

• BarHomogeneous: selects the barrier homogeneous algorithm

• BarOrder: determines the fill reducing ordering strategy

• Crossover: controls barrier crossover

• CrossoverBasis: controls initial crossover basis construction

• QCPDual: enables dual variable computation for continuous QCP models

• MIP: affect the MIP algorithms

• BranchDir: controls the branching node selection

• ConcurrentMIP: enables concurrent MIP optimization

• ConcurrentJobs: enables distributed concurrent optimization

• DegenMoves: limit degenerate simplex moves

• Disconnected: controls the disconnected component strategy
  Heuristics: controls the amount of time spent in MIP heuristics