ml4cps.automata.base module

The module provides a class Automaton which inherits CPSComponent and implements the dynamics of different kinds of automata.

Authors: - Nemanja Hranisavljevic, hranisan@hsu-hh.de, nemanja@ai4cps.com - Tom Westermann, tom.westermann@hsu-hh.de, tom@ai4cps.com

class ml4cps.automata.base.Automaton(states: list = None, transitions: list = None, unknown_state: str = 'raise', id='', initial_q=(), final_q=(), super_states=(), decision_states=(), **kwargs)

Bases: CPSComponent

Automaton class is the main class for modeling various kinds of hybrid systems.

property Sigma

accepts(string, return_states=False)

add_final_state(states)

Add final state(s) of the automaton.

param states:

States to add.

add_initial_state(states)

Add initial state(s) of the automaton.

param states:

States to add.

add_single_transition(s, d, e, timing=None)

add_state(new_state, **kwargs)

Add state to the automaton.

param new_state:

State to be added.

add_state_data(s: str, d: object)

Add state data to a state s the automaton.

param s:

state

param d:

data to be added to s

return:

add_states_from(new_state, **kwargs)

Add multiple states to the automaton.

param new_state:

States to be added.

add_transition(s, d, e, **other)

Add multiple transition.

param list_of_tuples:

List of transitions in the form (source_state, destination_state, event, …<unused>…).

add_transitions_from(list_of_tuples, **other)

Add multiple transition.

param list_of_tuples:

List of transitions in the form (source_state, destination_state, event, …<unused>…).

determinize(s, state_index, verbose=False)

discrete_event_dynamics(q, xt, xk, p) → tuple

property discrete_state

property discrete_states

flow(q, p, x, u)

Flow equation gives derivative of the continuous variables.

param q:

Current discrete state of the model.

param p:

Stochastic parameters generated on entry to state current_q.

param x:

Current continuous state of the model.

param u:

Calculated internal i signals.

return:

Derivative of the continuous state x.

generate(number_of_sequences=1, return_states=False, prob_to_accept=0.5)

get_alternatives(state, system_state)

get_num_in(q)

Returns the number of in transitions of state q in the automaton. :return: number of transitions.

get_num_out(q)

Returns the number of out transitions of state q in the automaton. :return: number of transitions.

get_transition(s, d=None, e=None, if_more_than_one='raise')

Get all transitions with source state s, destination state __d. In case when e is provided, the returned list contains transitions where event is e. :param if_more_than_one: :param s: Source state. :param d: Destination state. :param e: Event. :return:

get_transitions()

guards(q, x)

in_transitions(s, event=None)

Get all incoming transitions of state s.

param s:

return:

inv(t, q, x, y, z, p)

Invariants.

param t:

param q:

param x:

param y:

param z:

param p:

is_decision(state, overall_state)

is_deterministic()

is_state(q)

is_transition(s, d, e)

Check if a transition (s,d,e) exists in the automaton.

param s:

Source.

param d:

Destination.

param e:

Event.

return:

learn_transition_probabilities()

merge(q1, q2)

If two states are compatible, they are merged with the function merge. The transitions of the automaton, the in- and outdegree of the states and the number of transitions happening are adjusted.

property num_events

property num_modes

Returns the number of modes in the automaton. :return: number of states.

num_occur(tr)

num_timings()

property num_transitions

Returns the number of transitions in the automaton. :return: number of transitions.

out_transitions(s, event=None)

Get all outgoing transitions of state s.

param event:

param s:

return:

predict_state(data_collection, time_col_name=None, discr_col_names=None)

print_state(v)

Prints outgoing transitions of a state v.

Parameters:

v (state)

Returns:

Description of the outgoing transitions of the state.

Return type:

String

read_event(t, e, clear_p=False, keep_p=None, **kwargs)

remove_rare_transitions(min_p=0, min_num=0, keep_from_initial=False, keep_states=False, keep=None)

remove_state(s)

remove_transition(source, dest)

Remove the transition(s) from source to dest. :param source: :param dest: :return:

rename_events(prefix='e_')

Rename events to become e_0, e_1… The old id is stored in the field ‘old_symbol’ of the state data.

sample_initial()

property state

Automata discrete state is uni-variate. :return:

state_is_deterministic(q)

step(q, x0, t, u)

Simulates one time step of continuous behavior from t to t+dt. Underlying function is solve_ivp with method is ‘RK23’.

param x0:

Initial state vector.

param t:

Time at start of the step simulation.

param u:

Arguments passed…..

return:

Time t+dt, value of state at t+dt

timed_event(q, xc, xd)

Calculates if and when the next time event will happen and the new state values. :param t: Current time. :param q: Discrete-event part of the state. :param xc: Time-continuous part of the state. :param xd: Time-discrete part of the state. :return: Time delay of the timed event, new state value.

property transitions

try_merge_states(state1, state2, try_fun=None)

Merge state2 into state1 and update transitions.

update_timing_boundaries(source, destination, event, newTiming)