API Documentation#

class model.HybridModel(ABM: networkx.classes.graph.Graph)#

Create a ABM/FCM hybrid model for simulation based on a created the graph Each node of the ABM graph will have an ‘FCM’ attribute that points to an FCM graph. Each node of an FCM graph will have a ‘val’ attribute that has the value of that node.

Parameters

ABM (networkx.Graph) – The ABM/FCM graph created using networkx

generate_communities(community_algo)#

Partition the network into communities

Parameters

community_algo (func) – the community algorithm to be used

get_neighbors(node: int) → list#

Get neighbors for node in an adjacency matrix representation of a graph

Parameters

node (int) – the node for which neighbors are fetched

Returns

a list of neighbors of the requested node

Return type

list

loadNewValues() → None#

Sets FCM values to future values for each FCM in each agent infrom numba import cuda the simulation

runFCM(focus, threshold, max_it=100)#

Runs fuzzy cognitive map for all agents

Parameters

  • focus (list) – list of focus nodes

  • threshold (list) – a list of thresholds corresponding to each focus node

  • max_it (int, optional) – the maximum number of iterations for each agent, defaults to 100

run_parallel(focus, threshold, iters, func, args, steps=20, with_community=False, community_algo=None)#

Run the interaction function and stabelize each FCM in parallel for each step

Parameters

  • focus (list) – list concept names to stabilize

  • threshold (list) – stabilization thresholds for corresponding focus nodes

  • iters (int) – maximum amount of times to iterate

  • func (function) – interaction function between agents

  • args (list) – list of arguments for interaction function

  • steps (int) – number of steps of the simulation

Returns

average values for each focus concept

Return type

Dict

run_parallel_community(focus, threshold, iters, func, args, steps, community_algo)#

Run the interaction function and stabelize each FCM in parallel for each step

Parameters

  • focus (list) – list concept names to stabilize

  • threshold (list) – stabilization thresholds for corresponding focus nodes

  • iters (int) – maximum amount of times to iterate

  • func (function) – interaction function between agents

  • args (list) – list of arguments for interaction function

  • steps (int) – number of steps of the simulation

Returns

average values for each focus concept

Return type

Dict

run_serial(focus, threshold, iters, func, argv, steps=20) → float#

Run the model serially on CPU

Parameters

  • focus (list(Float64)) – a list of focus nodes

  • threshold (list(Float64)) – a list of thresholds for each focus nodes

  • iters (int) – the maximum number of iterations when running FCM for each agent

  • func (function) – the function for interaction between agents

  • argv (list) – a list of arguments for the interaction function

  • steps (int, optional) – the number of steps the model will be run, defaults to 20

Returns

the average values of the focus concepts of all agents of after the simulation

Return type

dict

transformNetwork(G: networkx.classes.graph.Graph)#

Use the networkx.Graph provided to extract information, including ABM and FCM adjacency matrices, a dictionary that maps FCM lables to their indices in the FCM adjacency matrix, along with other information

Parameters

G (networkx.Graph) – networkx graph to be transformed

model.create_FCM_file(filename) → networkx.classes.digraph.DiGraph#

Create FCM graph for an agent

Parameters

filename (string) – the path to the file for FCM model

Returns

a directed graph based on the provided file

Return type

networkx.DiGraph

model.create_graph(agent_count, graph_type, filename)#

Create ABM graph

Parameters

  • agent_count (int) – the number of agents in the model

  • graph_type (string) – the graph type to be created (watts, sf, newman, or barabasi)

  • filename (string) – the path to a text file that has the edges for the FCM graph

Returns

a graph for the ABM/FCM hybrid model

Return type

networkx.Graph

model.generate_model(agent_count, filename, graph_type)#

Create a general model

Parameters

  • agent_count (int) – number of agents in the model

  • filename (string) – the path to a file that stores the edges for the FCM

  • graph_type (string) – the type of graphs, pick from watts, sf, newman, and barabasi

Returns

a HybridModel

Return type

HybridModel