`mango_library.negotiation.cohda`¶

Module for distributed real power planning with COHDA. Contains roles, which integrate COHDA in the negotiation system and the core COHDA-decider together with its model.

class mango_library.negotiation.cohda.COHDA(weights, schedule_provider, is_local_acceptable, part_id)¶

COHDA-decider

decide(content: mango_library.negotiation.cohda.CohdaMessage) → Tuple[mango_library.negotiation.cohda.WorkingMemory, mango_library.negotiation.cohda.WorkingMemory]¶

Execute the COHDA decision process.

Parameters: content – the incoming COHDA message
Returns: old and new working memory

objective_function(candidate, target_schedule)¶

Objective function of COHDA. Calculates the negative sum of all deviations of the candidate to the target schedule

Parameters

candidate – candidate schedule
target_schedule – target schedule

Returns

negative sum of all deviations

class mango_library.negotiation.cohda.COHDARole(schedules_provider, weights, local_acceptable_func)¶

Negotiation role for COHDA.

create_cohda(part_id: int)¶

Create an instance of the COHDA-decider.

Parameters: part_id – participant id
Returns: COHDA

handle(message, assignment: mango_library.coalition.core.CoalitionAssignment, negotiation: mango_library.negotiation.core.Negotiation, meta: Dict[str, Any])¶

Handle the message and execute the specific negotiation step.

Parameters

message – the message
assignment – the assignment the negotiations is in
negotiation – the negotiation model
meta – meta data

class mango_library.negotiation.cohda.CohdaMessage(working_memory: mango_library.negotiation.cohda.WorkingMemory)¶

Message for a COHDa negotiation. Contains the current working memory of an agent.

property working_memory: mango_library.negotiation.cohda.WorkingMemory¶

Return the working memory of the sender agent

Returns: the working memory of the sender

class mango_library.negotiation.cohda.CohdaNegotiationStarterRole(target_schedule, coalition_model_matcher=None, coalition_uuid=None)¶: Convenience role for starting a COHDA negotiation with simply providing a target schedule

class mango_library.negotiation.cohda.ScheduleSelection(schedule: numpy.array, counter: int)¶

A selection of a specific schedule

property counter: int¶

The counter of the selection

Returns: the counter

property schedule: numpy.array¶

The schedule as np.array

Returns: scheduke

class mango_library.negotiation.cohda.SolutionCandidate(agent_id: int, candidate: Dict[int, numpy.array])¶

Model for a solution candidate in COHDA.

property agent_id: str¶

Return the agent id

Returns: agent id

property candidate: Dict[int, numpy.array]¶

Return the candidate schedule map (part_id -> schedule)

Returns: map part_id -> schedule

class mango_library.negotiation.cohda.WorkingMemory(target_schedule, system_config: Dict[str, mango_library.negotiation.cohda.ScheduleSelection], solution_candidate: mango_library.negotiation.cohda.SolutionCandidate)¶

Working memory of a COHDA agent

property solution_candidate: mango_library.negotiation.cohda.SolutionCandidate¶

The current best known solution candidate for the planning

Returns: the solution candidate

property system_config: Dict[str, mango_library.negotiation.cohda.ScheduleSelection]¶

Return the system config as map part_id -> selection

Returns: the believed system state

property target_schedule¶

Return the target schedule

Returns: the target schedule

mango_library.negotiation.cohda¶

`mango_library.negotiation.cohda`¶