Iterative Pricing Exercise

[mhumpher]

I load charges/premiums as new cells into one of my spaces. However, as part of pricing exercises we might want to iteratively test different sets of charges/premiums. Is there a way to replace or update input cells?

[fumitoh]

You can assign input values to cells by [] =

For example,

import modelx as mx
from modelx import defcells

@defcells
def pv_premiums():
    return 0 # dummy

@defcells
def pv_claims():
    return 3

@defcells
def net_pv():
    return pv_premiums() - pv_claims()

and

>>> pv_premiums[()] = 4

>>> net_pv()
1

>>> pv_premiums[()] = 3

>>> net_pv()
0

You can write a loop in Python to iterate the assignment operation above. but it's more recommended to refer to a Reference instead of directly assigning values to pv_premiums:

@defcells
def pv_premiums():
    return pv_prem_input

>>> mx.Model1.Space1.pv_prem_input = 4

>>> net_pv()
1

>>> mx.Model1.Space1.pv_prem_input = 3

>>> net_pv()
0

[fumitoh]

Or, you can even write a solver using minimize_scalar from scipy.

from scipy.optimize import minimize_scalar  

def solv_premium(space):
   
   def solv_prem_inner(x):
       space.pv_prem_input  = x
       return space.net_pv()**2
   
   res = minimize_scalar(solv_prem_inner)
   return res.x

>>> solv_premium(mx.Model1.Space1)
3.0

>>> mx.Model1.Space1.pv_prem_input
2.99999995559

Furture more, you can write the solver code above in a *.py file and save it within the model using new_module.