Files in the folder (tested on Julia v1.8 and Python 3.9):

• test_ReverseDiff.jl : For testing the automatic differentiation packages in julia (read this first).

• GraphUtil.jl : Utility functions for read/write graphs, and various graph representations.

• SIR_xxx.jl : DMP + optimization on the usual SIR model

• mSIR_xxx.jl : DMP + optimization on a modified SIR model

• data_analysis.py : Visualize the result of "SIR_DMP_opt_sigma0.jl"

• artificial_networks : Directory for some networks

• results : Directory to dump results

Spreading process models considered:

• SIR : the usual discrete time SIR model, used in [A. Y. Lokhov et al., Phys. Rev. E 91, 012811, 2015].
• mSIR : the modified SIR model, used in [Lokhov and Saad PNAS2017].

Optimization method and tricks:

• Basic method: gradient ascent + (optional) backtracking line search
• Some reparametrization tricks are used to enforce the total resource constraint; see [Lokhov and Saad PNAS2017]
• Barrier functions are used to enforce some inequality constraints