Task-Assignment-Hypergraph-Discovery

Here you can find the code for the paper "Assigning Entities to Teams as a Hypergraph Discovery Problem"

Install Required Packages

pip install -r dependency.txt

Set the config file in configs directory and run "run.py" with the specified config file.

Parameters/Configs

Mode Parameters

• data: APS/MAG
• Algorithm: CSA/CSA_Bipartite/Greedy/GreedyBipartite/RandomGreedy/RandomGreedyBipartite

Folder Paths

• logging_path: logging path
• folder_path: data folder path
• res_path: result saving path

Optimization Parameters

• epoch: number of optimization rounds
• Nt: number of optimization rounds for each temperature
• tol: convergence tolerace
• temp: initial temperature
• temp_decay: temperature decay rate
• nremove: number of removed assignments in each perturbation
• nswap: number of assignment swaps in each perturbation
• sym: symmetric algebraic connectivity: default False
• pen, pen1, pen2, pen3: coefficients of the penalty terms
• return: probability of going back to the best found solution at each round
• pack: energy pack size

Robustness

To evaluate the solution robustness tp node failures, run "robustness.py" file with the config specified in "Robustness.json"

Parameters/Configs for Robustness

• set any of APS/Bipartite_APS/Best_APS/Initial_APS/Greedy_APS/MAG/Bipartite_MAG/Best_MAG/Greedy_MAG/Initial_MAG to True to evaluate the corresponding solution

• Natt: number of removed (attacked) nodes

• niter: number of times the solution is evaluated