Markets

The data structures provided here are a basis for agent-based market simulations. For example, they could be used in financial market models or labor market models.

Market Registry

class sfctools.datastructs.market_registry.MarketRegistry(owner)

Bases: object

Data structure to store remaining and fulfilled demand of an agent at a market

DEMAND:

total

fulfilled

remaining

good1

=total-fulfilled

good2

good3

SUPPLY:

total

fulfilled

remaining

good1

=total-fulfilled

good2

good3
Parameters

owner – agent instance who owns this data structure

reset(kind, good)

reset all entries to zero

Parameters

  • kind – ‘demand’ or ‘supply’

  • good – name of good

set_current_update(kind, good, q)

” update fulfillment of demand or supply

Parameters

  • kind – ‘demand’ or ‘supply’

  • good – name of good

  • q – new entry value

set_target_update(kind, good, new_value)

set up the toftal demand/supply for a certain good

Parameters

  • kind – ‘demand’ or ‘supply’

  • good – name of good

  • q – new entry value

to_string()

converts data structure to human-readable string format

class sfctools.datastructs.market_registry.RegistryEntry

Bases: object

A helper data structure for MarketRegistry

Market Matching

class sfctools.bottomup.matching.MarketMatching(demand_list=None, supply_list=None)

Bases: object

Meta class for market matching. This is a basic framework for a list of demanders and suppliers who are matched decentrally and registered centrally. This can be thought of as a graph (nodes and edges) of agents.

constructor for market matching

Parameters

  • demand_list – list of demand agents or None (default, creates empty list)

  • supply_list – list of supply agents or None (default)

add_demander(agent)

Add demand agent as node in the graph

Parameters

agent – instance to add on demand side

add_supplier(agent)

Add supply agent as node in the graph

Parameters

agent – instance to add on demand side

clear()

Removes all edges from the current graph.

clear_all()

Do a complete reset (remove edges and nodes, reset to blank greenfield market.

property demand_list

Get list of demand agents. Filters demanders who are not bankrupt

Returns

list of agents

get_demand_data(agent)

get all demanders + info for agent

Parameters

agent – reference agent

Returns

dict containing data

get_demanders_from(agent) → list

get list of agents demanding from agent

Parameters

agent – reference agent

Returns

list of agents

get_suppliers_of(agent) → list

get list of agents supplying to agent

Parameters

agent – reference agent

Returns

list of agents

get_supply_data(agent)

get all suppliers + info for agent

Parameters

agent – reference agent

Returns

dict containing data

get_value(agent_from, agent_to)

link a connection between demand and supply with weight val

Parameters

  • supply_agent – agent instance, supply

  • demand_agent – agent instance, demand

  • val – a value that is stored in the data dict for this link

plot(fname=None, show_figure=True)

plot the current matching situation as a network graph :param fname: filename (optional) if filename is given, no plot will be shown and a png will be saved to a file. :param show_figure: default True, show the figure in a plot window

plot_weighted(my_layout=None, arrows=False, fname=None)

plot the current matching situation as a network graph, with edge thickness matching the weights Suppliers re plotted in red, demanders in blue.

Parameters

  • my_layout – networkx graph layout

  • arrows – show the arrows

  • fname – if None, plot is shown in window. Else, file is stored to fname

abstract rematch()

rematch the supply and demand agents, i.e. update the matching map. Has to be overridden as it is abstract.

remove_demander(agent)

remove a demander

remove_supplier(agent)

remove a supplier

property supply_list

Get list of supply agents. Filters suppliers who are not bankrupt

Returns

list of agents

unlink a connection

Parameters

  • supply_agent – agent instance, supply

  • demand_agent – agent instance, demand