Является результатом Глобального MINLP Solver, глобального оптимума - PullRequest
Новая
       52

Является результатом Глобального MINLP Solver, глобального оптимума

1 голос
/ 19 апреля 2020

Я использовал Pyomo и Couenne Solver, который является глобальным решателем для MINLP. Но результат не является глобальным оптимальным! Возможно ли, что результат глобального решения не будет глобальным оптимумом? В чем проблема?

Журнал решателя: 

Solver command line: ['C:\\Users\\~\\couenne.exe', 'C:\\Users\\~\\AppData\\Local\\Temp\\tmpv6y0s3cc.pyomo.nl', '-AMPL']
couenne: 
ANALYSIS TEST: Problem size before reformulation: 16 variables (6 integer), 37 constraints.
Problem size after  reformulation: 54 variables (6 integer), 37 constraints.
*****************************************************************************
This program contains Ipopt, a library for large-scale nonlinear optimization.

 Ipopt is released as open source code under the Common Public License (CPL).
         For more information visit http://projects.coin-or.org/Ipopt ******************************************************************************
NOTE: You are using Ipopt by default with the MUMPS linear solver.

      Other linear solvers might be more efficient (see Ipopt documentation).
Cbc0031I 8 added rows had average density of 2

Cbc0013I At root node, 8 cuts changed objective from -20.3231 to -20.235 in 10 passes

Cbc0014I Cut generator 0 (Couenne convexifier cuts) - 48 row cuts average 2.0 elements, 2 column cuts (10 active)

Cbc0010I After 0 nodes, 0 on tree, 1e+050 best solution, best possible -20.235 (0.16 seconds)

Cbc0004I Integer solution of -10.0826 found after 831 iterations and 63 nodes (0.81 seconds)

Cbc0010I After 100 nodes, 38 on tree, -10.0826 best solution, best possible -15.9137 (0.89 seconds)

Cbc0004I Integer solution of -10.5219 found after 1264 iterations and 114 nodes (0.91 seconds)

Cbc0010I After 200 nodes, 26 on tree, -10.5219 best solution, best possible -11.899 (1.44 seconds)

Cbc0010I After 300 nodes, 29 on tree, -10.5219 best solution, best possible -10.9884 (1.61 seconds)

Cbc0010I After 400 nodes, 13 on tree, -10.5219 best solution, best possible -10.5235 (1.72 seconds)

Cbc0001I Search completed - best objective -10.52192111756512, took 3256 iterations and 479 nodes (1.80 seconds)

Cbc0035I Maximum depth 36, 0 variables fixed on reduced cost

 Couenne convexifier cuts was tried 939 times and created 3190 cuts of which 548 were active after adding rounds of cuts



    "Finished"
...