mean_neighbor_cost value in history Tuple of optimizers

[mmfarrugia]

I noticed while extending pyswarms for epsilon constrained optimization that the value passed to mean_neighbor_cost in the SwarmOptimizer.ToHistory Tuple (the hist tuple which is then passed to SwarmOptimizer._populate_history method as an argument) is just self.swarm.best_cost. Did I misunderstand the way that this works since swarm.best_cost comes from the topology's implementation of compute_gbest method, or should this be a mean of the neighboring costs not a max/best global cost? I noticed that it is indeed the mean in local_best so I figured that this is either unintentional although the code works or I need to better understand the implementation and how mean_neighbor_cost is used because it makes sense that there would be this kind of difference but I couldn't seem to track down where it is used in the code. I understand that it is then appended to mean_neighbor_history, but I couldn't find any references to this variable that used it. Thank you for your help, I hope to contribute a constrained implementation of the pyswarms library soon based off of Takahama et al 2005 epsilon approach.

It is on line 261 of general_optimizer.py, line 221 on global_best.py, line 249 of local_best.py in pyswarms/single folder.