Integer Optimization Controls

class ecyglpki.IntOptControls

The integer optimization solver control parameter object

>>> r = IntOptControls()
binarize

Whether to binarize integer variables, a bool

This option is only used if presolve is True.

br_tech

The branching technique, a str

The possible values are

  • 'first_fracvar': first fractional variable
  • 'last_fracvar': last fractional variable
  • 'most_fracvar': most fractional variable
  • 'Driebeek-Tomlin': heuristic by Driebeek & Tomlin
  • 'hybrid_peudocost': hybrid pseudocost heuristic
bt_tech

The backtracking technique, a str

The possible values are

  • 'depth': depth first search
  • 'breadth': breadth first search
  • 'bound': best local bound
  • 'projection': best projection heuristic
cb_func

Callback routine

cb_info

Transit pointer passed to the routine cb_func

cb_size

Number of extra bytes allocated for each search tree node, an int

Up to 256 bytes can be allocated for each node of the branch-and-bound tree to store application-specific data. On creating a node these bytes are initialized by binary zeros.

>>> r.cb_size  # the GLPK default
0
>>> r.cb_size = 128
>>> r.cb_size
128
clq_cuts

Whether to generate generate clique cuts, a bool

cov_cuts

Whether to generate mixed cover cuts, a bool

fp_heur

Whether to apply the feasibility pump heuristic, a bool

gmi_cuts

Whether to generate Gomory’s mixed integer cuts, a bool

mip_gap

The relative MIP-gap tolerance, a Real number

The search stops once the relative MIP-gap falls below this value.

mir_cuts

Whether to generate mixed integer rounding cuts, a bool

msg_lev

The message level, a str

The possible values are

  • 'no': no output
  • 'warnerror': warnings and errors only
  • 'normal': normal output
  • 'full': normal output and informational messages
>>> r.msg_lev  # the GLPK default
'full'
>>> r.msg_lev = 'no'
>>> r.msg_lev
'no'
out_dly

Output delay [ms] of current LP relaxation solution, an int

out_frq

Output frequency [ms] of informational messages, an int

pp_tech

The preprocessing technique, a str

The possible values are

  • 'none': disable preprocessing
  • 'root': preprocessing only on the root level
  • 'all': preprocessing on all levels
presolve

Whether to use the MIP presolver, a bool

Using the MIP presolver may simplify the problem

ps_heur

Whether to apply the proximity search heuristic, a bool

ps_tm_lim

Time limit [ms] for the proximity earch heuristic, an int

tm_lim

Time limit [ms], an int

tol_int

Abs. tolerance for LP solution integer feasibility, a Real number

This is the absolute tolerance used to check if the optimal solution to the current LP relaxation is integer feasible.

tol_obj

Rel. tolerance of LP objective optimality, a Real number

This is the relative tolerance used to check if the objective value in the optimal solution to the current LP relaxation is not better than in the best known integer feasible solution.

Previous topic

Interior Point Controls

Next topic

Branching Tree

This Page