system run.py

for iteration in range(iters * batch_size):
    # Create a graph with random weights
    G = nx.complete_graph(num_nodes, nx.DiGraph())
    # ... add random weights ...

    # Give an edge to a sample of agents (not all are going to be playing)
    players = random.sample(agents, len(G.edges))

    # Calculate the shortest path
    path = nx.shortest_path(G, source=source, target=sink, weight="price")

    # Pay each player on the shortest path
    for edge in get_edge_list(path):
        pay_player_owning_edge(edge)

    # Once a batch of runs is done, prune players and add new ones
    if iteration % batch_size == 0:
        prune_poor_agents()
        clone_wealthy_agents()
        add_new_agents()