MCFI: Minimum Cost Flow Interpolation¶

Class Description¶

class paminco.algo.mcfi.MCFI(net: paminco.net.network.Network, name=None, callback=None, use_simple_timer: bool = True, optim=None, kw_optim=None, **kwargs)[source]¶

Minimum Cost Flow Interpolation.

Computes min costs flows for breakpoints. For parameters in-between the breakpoints, a min cost flow is then interpolated.

Parameters

netNetwork

Network to find min cost flows for. A min cost flow f for some param is found by:

minimize net.cost(f)
s.t. net.Gamma.dot(f) = net.demand(param).

namestr, optional

Name of solver.

callbackcallable, or list of callable, optional

Callbacks called during initialization and run method. Can be used for debugging and timing.

use_simple_timerbool, default=True

Whether to time MCFI. If True, timestamps for intializtion and every iteration will be saved to attribute timer.

kw_optimkeyword arguments, optional

Further arguments passed to min cost flow optimizer.

kwargskeyword arguments, optional

Further parameters for MCFI, see MCFIConfig.

See also

MCFIConfig: All settings for MCFI.
paminco.optim.fw_net.NetworkFW: Optimizer to find min cost flows for a given parameter.

Examples

Find user equilibrium (UE) for various specified demand factors:

>>> import paminco
>>> net = paminco.net.load_sioux()
>>> net.integrate_cost()  # integrate to calc UE
>>> mcfi = MCFI(net)
>>> demand_fac = [0.2, 0.5, 0.6]
>>> mcfi.run(param=demand_fac)
>>> mcfi.cost_at(demand_fac).round(2)
array([ 638722.63, 1673514.68, 2062773.13])

Attributes

configMCFIConfig: Settings of MCFI.
networkNetwork: Get network obj.
namestr: Get name of object.
param_solutionParametricSolution: Get ParametricSolution object.

Settings¶

`MCFIConfig`(**kw)	Settings for the Marginal Cost Flow Interpolator.
`AdaptiveMethod`(value)	Enum defining the step rule used in the MCFI algorithm.

Attributes¶

`MCFI.network`	Get network obj.
`MCFI.name`	Get name of object.
`MCFI.config`	Settings of MCFI.
`MCFI.param_solution`	Get ParametricSolution object.

Methods¶

MCFI.run([param, callback, kw_optim])

Run MCFI.