I’m a newbie researcher working on model-based genetic algorithms, mainly linkage learning in both discrete and continuous spaces, using data modeling. I would like to ask you about Ising Spin Glass (ISG) problem in the context of optimization. I’ve seen a lot of papers running benchmarks on this particular problem. I’ve googled it, checked wiki but eventually understood nothing :/
- What is an intuitive definition of this problem, making it well suited for optimization?
- When referring to ISG problem in papers, which specific model researchers have in mind?
- How to formulate this problem in the context of optimization? (encoding, fitness evaluation)
I don’t ask nor expect a ready solution, some general direction would suffice as I feel kinda lost.
I’ve found some python repo on github. It calculates energy in a following way:
import numpy as np def totalEnergy(J, configuration): """ calculate the energy for a given configuration. Parameters: >>> J: nSpin*nSpin matrix, J[i,j] is the interation between spin i and j. >>> configuration: 1*nSpin vector, configuration[i] stores the spin at site i. Returns: the total energy of this configuration. """ # 0.5 for double counting return 0.5*np.dot(configuration,np.dot(J,configuration)) ```