Issues regarding constraint's setting via mystic package

Question

I am trying to set constraints to perform optimization with the help of mystic package.

I executed the code in Spyder and PyCharm IDE. In both cases kernel crashed. When I have only 2 line string 'simplify'-method works fine. Trying to set greater than 2 lines constraints leads to kernel's death.

import mystic.symbolic as ms

equation = """
x0*7.2 + x1*1.9 + x2*35.7 + x3*4.1 + x4*23.0 + x5*19.2 + x6*0.0 + x7*0.0 + x8*0.0 + x9*0.0 + x10*10.1 + x11*10.8 + x12*11.3 + x13*14.8 + x14*78.6 + x15*5.8 + x16*3.4 + x17*1.8 + x18*3.4 + x19*1.0 + x20*2.9 + x21*0.0 + x22*0.0 + x23*0.0 + x24*0.0 + x25*0.0 + x26*0.0 + x27*0.0 + x28*0.0 + x29*0.0 >= 27.0
x0*7.2 + x1*1.9 + x2*35.9 + x3*4.1 + x4*23.0 + x5*19.2 + x6*0.0 + x7*0.0 + x8*0.0 + x9*0.0 + x10*9.8 + x11*10.9 + x12*27.1 + x13*15.0 + x14*78.7 + x15*5.8 + x16*3.4 + x17*1.8 + x18*3.4 + x19*1.0 + x20*2.9 + x21*0.0 + x22*0.0 + x23*0.0 + x24*0.0 + x25*0.0 + x26*0.0 + x27*0.0 + x28*0.0 + x29*0.0 >= 29.0
x0*7.2 + x1*1.9 + x2*35.9 + x3*4.1 + x4*23.0 + x5*19.2 + x6*0.0 + x7*0.0 + x8*0.0 + x9*0.0 + x10*9.8 + x11*10.9 + x12*27.1 + x13*15.0 + x14*78.7 + x15*5.8 + x16*3.4 + x17*1.8 + x18*3.4 + x19*1.0 + x20*2.9 + x21*0.0 + x22*0.0 + x23*0.0 + x24*0.0 + x25*0.0 + x26*0.0 + x27*0.0 + x28*0.0 + x29*0.0 >= 29.0
x0*0.0 + x1*0.0 + x2*0.0 + x3*0.0 + x4*0.0 + x5*0.0 + x6*27.9 + x7*73.0 + x8*230.9 + x9*107.3 + x10*0.0 + x11*0.0 + x12*0.0 + x13*0.0 + x14*0.0 + x15*0.0 + x16*0.0 + x17*0.0 + x18*0.0 + x19*0.0 + x20*0.0 + x21*13.7 + x22*63.7 + x23*29.5 + x24*7.2 + x25*24.3 + x26*9.6 + x27*142.1 + x28*10.5 + x29*41.0 >= 420.0
x0*0.0 + x1*0.0 + x2*0.0 + x3*0.0 + x4*0.0 + x5*0.0 + x6*27.2 + x7*72.9 + x8*88.5 + x9*107.3 + x10*0.0 + x11*0.0 + x12*0.0 + x13*0.0 + x14*0.0 + x15*0.0 + x16*0.0 + x17*0.0 + x18*0.0 + x19*0.0 + x20*0.0 + x21*13.7 + x22*63.7 + x23*29.5 + x24*7.2 + x25*23.6 + x26*9.6 + x27*142.1 + x28*10.3 + x29*37.0 >= 420.0
"""

eqn = ms.simplify(equation)

score 0 · Answer 1 · answered May 01 '19 at 01:34

You are probably just hitting your memory limit, or something like that. I'd suggest to turn off iterating through all solutions, and turn on cycling the variable solved for.

>>> import mystic.symbolic as ms
>>> equation = """
... x0*7.2 + x1*1.9 + x2*35.7 + x3*4.1 + x4*23.0 + x5*19.2 + x6*0.0 + x7*0.0 + x8*0.0 + x9*0.0 + x10*10.1 + x11*10.8 + x12*11.3 + x13*14.8 + x14*78.6 + x15*5.8 + x16*3.4 + x17*1.8 + x18*3.4 + x19*1.0 + x20*2.9 + x21*0.0 + x22*0.0 + x23*0.0 + x24*0.0 + x25*0.0 + x26*0.0 + x27*0.0 + x28*0.0 + x29*0.0 >= 27.0
... x0*7.2 + x1*1.9 + x2*35.9 + x3*4.1 + x4*23.0 + x5*19.2 + x6*0.0 + x7*0.0 + x8*0.0 + x9*0.0 + x10*9.8 + x11*10.9 + x12*27.1 + x13*15.0 + x14*78.7 + x15*5.8 + x16*3.4 + x17*1.8 + x18*3.4 + x19*1.0 + x20*2.9 + x21*0.0 + x22*0.0 + x23*0.0 + x24*0.0 + x25*0.0 + x26*0.0 + x27*0.0 + x28*0.0 + x29*0.0 >= 29.0
... x0*7.2 + x1*1.9 + x2*35.9 + x3*4.1 + x4*23.0 + x5*19.2 + x6*0.0 + x7*0.0 + x8*0.0 + x9*0.0 + x10*9.8 + x11*10.9 + x12*27.1 + x13*15.0 + x14*78.7 + x15*5.8 + x16*3.4 + x17*1.8 + x18*3.4 + x19*1.0 + x20*2.9 + x21*0.0 + x22*0.0 + x23*0.0 + x24*0.0 + x25*0.0 + x26*0.0 + x27*0.0 + x28*0.0 + x29*0.0 >= 29.0
... x0*0.0 + x1*0.0 + x2*0.0 + x3*0.0 + x4*0.0 + x5*0.0 + x6*27.9 + x7*73.0 + x8*230.9 + x9*107.3 + x10*0.0 + x11*0.0 + x12*0.0 + x13*0.0 + x14*0.0 + x15*0.0 + x16*0.0 + x17*0.0 + x18*0.0 + x19*0.0 + x20*0.0 + x21*13.7 + x22*63.7 + x23*29.5 + x24*7.2 + x25*24.3 + x26*9.6 + x27*142.1 + x28*10.5 + x29*41.0 >= 420.0
... x0*0.0 + x1*0.0 + x2*0.0 + x3*0.0 + x4*0.0 + x5*0.0 + x6*27.2 + x7*72.9 + x8*88.5 + x9*107.3 + x10*0.0 + x11*0.0 + x12*0.0 + x13*0.0 + x14*0.0 + x15*0.0 + x16*0.0 + x17*0.0 + x18*0.0 + x19*0.0 + x20*0.0 + x21*13.7 + x22*63.7 + x23*29.5 + x24*7.2 + x25*23.6 + x26*9.6 + x27*142.1 + x28*10.3 + x29*37.0 >= 420.0
... """
>>> eqns = ms.simplify(equation, all=False, cycle=True)
>>> print(eqns)
x4 >= -0.31304347826087*x0 - 0.0826086956521739*x1 - 0.426086956521739*x10 - 0.473913043478261*x11 - 1.17826086956522*x12 - 0.652173913043478*x13 - 3.42173913043478*x14 - 0.252173913043478*x15 - 0.147826086956522*x16 - 0.0782608695652174*x17 - 0.147826086956522*x18 - 0.0434782608695652*x19 - 1.56086956521739*x2 - 0.126086956521739*x20 - 0.178260869565217*x3 - 0.834782608695652*x5 + 1.26086956521739
x7 >= -0.187928669410151*x21 - 0.873799725651577*x22 - 0.404663923182442*x23 - 0.0987654320987654*x24 - 0.323731138545953*x25 - 0.131687242798354*x26 - 1.94924554183813*x27 - 0.141289437585734*x28 - 0.507544581618656*x29 - 0.373113854595336*x6 - 1.21399176954733*x8 - 1.4718792866941*x9 + 5.76131687242798
x0 >= -0.263888888888889*x1 - 1.40277777777778*x10 - 1.5*x11 - 1.56944444444444*x12 - 2.05555555555556*x13 - 10.9166666666667*x14 - 0.805555555555556*x15 - 0.472222222222222*x16 - 0.25*x17 - 0.472222222222222*x18 - 0.138888888888889*x19 - 4.95833333333333*x2 - 0.402777777777778*x20 - 0.569444444444444*x3 - 3.19444444444444*x4 - 2.66666666666667*x5 + 3.75
x6 >= -0.491039426523298*x21 - 2.2831541218638*x22 - 1.0573476702509*x23 - 0.258064516129032*x24 - 0.870967741935484*x25 - 0.344086021505376*x26 - 5.09318996415771*x27 - 0.376344086021505*x28 - 1.46953405017921*x29 - 2.61648745519713*x7 - 8.27598566308244*x8 - 3.84587813620072*x9 + 15.0537634408602
x3 >= -1.75609756097561*x0 - 0.463414634146341*x1 - 2.39024390243902*x10 - 2.65853658536585*x11 - 6.60975609756098*x12 - 3.65853658536585*x13 - 19.1951219512195*x14 - 1.41463414634146*x15 - 0.829268292682927*x16 - 0.439024390243902*x17 - 0.829268292682927*x18 - 0.24390243902439*x19 - 8.75609756097561*x2 - 0.707317073170732*x20 - 5.60975609756098*x4 - 4.68292682926829*x5 + 7.07317073170732
>>> 
>>> constrain = ms.generate_constraint(ms.generate_solvers(eqns))  
>>>
>>> constrain(list(range(-29,1)))
[-29, -28, -27, -26, 187.46956521739128, -24, 84.56076719225553, 85.54595336076824, -21, -20, -19, -18, -17, -16, -15, -14, -13, -12, -11, -10, -9, -8, -7, -6, -5, -4, -3, -2, -1, 0]
>>>

You may also want to look at and and or, which when used with the join keyword in generate_constraint, can be useful for modifying how the individual constraints work together. Also, if you have specific parameters you'd prefer are solved for, you can give them with the target keyword.

Anyway, the above should generate a constraint, useful within mystic's optimizers.

indeed, my memory usage skyrocketed and at the moment limit is reached, kernel crashed. Your suggestions worked for 5-line constraint mentioned in the title post. However, 42-line constraints still lead to crushing. I have tried to use ```all=False, cycle=True``` arguments in ```simplify``` as well as iterative appending of constraints simplified 1 by 1 to an empty list (for loop, list comprehension). All the methods ended with crashes. Don't know how to overcome this issue. — kimsergeo, May 01 '19 at 09:37
The thing that `simplify` does is essentially isolate one variable. My next suggestion was going to be to do a few equations at a time, and then concatenate the strings. There is, really, no need to do all the equations at once... and the `simplify` code can probably also be refactored so it doesn't do all equations at once either. When I get some time to, I'll look into reducing the memory footprint of the code. Can you report this as an issue on `mystic`'s GitHub page? Please also refer to this posting. — Mike McKerns, May 01 '19 at 14:26

Issues regarding constraint's setting via mystic package

1 Answers1

Linked