Currently solving a LP problem by GEKKO Library in python. The problem has like 8,000 variables. it runs fine with 462 variables if I increase the number of variables. The program shows an error "Exception: @error: Model Expression *** Error in syntax of function string: Invalid element: ,>=0"
Any comment will be helpful.
#Initialize Model
m = GEKKO()
#Set global options
m.options.solver = 3
m.options.IMODE = 3 #steady state optimization
#define parameter
Num_Cell= 463
H2 = m.Array(m.Var,Num_Cell,value = 0)
Demand = m.Const(value=20000000)
D = np.zeros(Num_Cell)
F = Hardwood_Sawlog.SumOfTotal
for i in range(Num_Cell):
H2[i].lower = 0
H2[i].upper = F[i]
D[i] = i
m.Equation(m.sum(H2[0:Num_Cell])==Demand)
m.Obj(np.dot(D,H2))
m.solve(disp=False,debug=True)
print('')
print(H2)
The problem is that the symbolic form of m.Obj(np.dot(D,H2))
exceeds the maximum length of 15,000 characters with the larger problem. Here is a version of your problem that produces an error:
from gekko import GEKKO
import numpy as np
m = GEKKO()
m.options.solver = 3
m.options.IMODE = 3
Num_Cell= 1000
H2 = m.Array(m.Var,Num_Cell,value = 0)
Demand = m.Const(value=20000000)
D = np.zeros(Num_Cell)
for i in range(Num_Cell):
H2[i].lower = 0
H2[i].upper = 100
D[i] = i
m.Equation(m.sum(H2[0:Num_Cell])==Demand)
m.Obj(np.dot(D,H2))
m.solve(disp=False,debug=True)
print('')
print(H2)
Here is an alternative version that overcomes the problem with np.dot()
constructing the long (>15,000 character) symbolic expression.
from gekko import GEKKO
import numpy as np
m = GEKKO()
m.options.solver = 1
m.options.IMODE = 3
Num_Cell= 1000
H2 = m.Array(m.Var,Num_Cell,value = 0)
Demand = m.Const(value=20000000)
D = np.zeros(Num_Cell)
for i in range(Num_Cell):
H2[i].lower = 0
H2[i].upper = 1e10
D[i] = i
m.Equation(m.sum(H2[0:Num_Cell])==Demand)
m.Minimize(m.sum([D[i]*H2[i] for i in range(Num_Cell)]))
m.solve(disp=True)
print('')
print(H2)
There are linear programming model building functions in gekko
that help with constructing large-scale LP problems with sparse matrices or large dense matrices:
An even more efficient way is to use the built-in model building functions .
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.