Implementing the Genetic Miner

docs/source/api.rst

@@ -90,6 +90,7 @@ Among *procedural process models*, ``pm4py`` currently supports:
   * :meth:`pm4py.discovery.discover_petri_net_inductive`; discovers a *Petri net* using the Inductive Miner algorithm.
   * :meth:`pm4py.discovery.discover_petri_net_heuristics`; discovers a *Petri net* using the Heuristics Miner algorithm.
   * :meth:`pm4py.discovery.discover_petri_net_ilp`; discovers a *Petri net* using the ILP Miner algorithm.
+  * :meth:`pm4py.discovery.discover_petri_net_genetic`; discovers a *Petri net* using the Genetic Miner algorithm.
   * :meth:`pm4py.discovery.discover_process_tree_inductive`; discovers a *process tree* using the Inductive Miner algorithm.
   * :meth:`pm4py.discovery.discover_bpmn_inductive`; discovers a *BPMN model* using the Inductive Miner algorithm.
   * :meth:`pm4py.discovery.discover_heuristics_net`; discovers a *heuristics net* using the Heuristics Miner algorithm.
@@ -466,6 +467,7 @@ List of Methods
    pm4py.discovery.discover_petri_net_inductive
    pm4py.discovery.discover_petri_net_heuristics
    pm4py.discovery.discover_petri_net_ilp
+   pm4py.discovery.discover_petri_net_genetic
    pm4py.discovery.discover_process_tree_inductive
    pm4py.discovery.discover_heuristics_net
    pm4py.discovery.derive_minimum_self_distance

examples/genetic_miner.py

@@ -0,0 +1,17 @@
+import pm4py
+import os
+import importlib.util
+from examples import examples_conf
+from pm4py.algo.discovery.genetic.algorithm import Parameters
+
+
+def execute_script():
+    log = pm4py.read_xes(os.path.join("..", "tests", "input_data", "running-example.xes"))
+    net, im, fm = pm4py.discover_petri_net_genetic(log, population_size = 20, generations = 30)
+
+    if importlib.util.find_spec("graphviz"):
+        pm4py.view_petri_net(net, im, fm, format=examples_conf.TARGET_IMG_FORMAT)
+
+
+if __name__ == "__main__":
+    execute_script()

pm4py/__init__.py

@@ -138,6 +138,7 @@
     discover_petri_net_ilp,
     discover_petri_net_heuristics,
     discover_petri_net_inductive,
+    discover_petri_net_genetic,
     discover_process_tree_inductive,
     discover_heuristics_net,
     discover_dfg,

pm4py/algo/discovery/__init__.py

@@ -33,6 +33,7 @@
         heuristics,
         ilp,
         inductive,
+        genetic,
         log_skeleton,
         minimum_self_distance,
         ocel,

pm4py/algo/discovery/genetic/__init__.py

@@ -0,0 +1,22 @@
+'''
+    PM4Py – A Process Mining Library for Python
+Copyright (C) 2026 Process Intelligence Solutions UG (haftungsbeschränkt)
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU Affero General Public License as
+published by the Free Software Foundation, either version 3 of the
+License, or any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU Affero General Public License for more details.
+
+You should have received a copy of the GNU Affero General Public License
+along with this program.  If not, see this software project's root or
+visit <https://www.gnu.org/licenses/>.
+
+Website: https://processintelligence.solutions
+Contact: info@processintelligence.solutions
+'''
+from pm4py.algo.discovery.genetic import variants, algorithm

pm4py/algo/discovery/genetic/algorithm.py

@@ -0,0 +1,77 @@
+'''
+    PM4Py – A Process Mining Library for Python
+Copyright (C) 2026 Process Intelligence Solutions UG (haftungsbeschränkt)
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU Affero General Public License as
+published by the Free Software Foundation, either version 3 of the
+License, or any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU Affero General Public License for more details.
+
+You should have received a copy of the GNU Affero General Public License
+along with this program.  If not, see this software project's root or
+visit <https://www.gnu.org/licenses/>.
+
+Website: https://processintelligence.solutions
+Contact: info@processintelligence.solutions
+'''
+from enum import Enum
+from pm4py.util import exec_utils
+from pm4py.algo.discovery.genetic.variants import classic
+from typing import Union, Optional, Dict, Any, Tuple
+from pm4py.objects.petri_net.obj import PetriNet, Marking
+from pm4py.objects.log.obj import EventLog, EventStream
+import pandas as pd
+from pm4py.util import constants
+
+
+class Parameters(Enum):
+    ACTIVITY_KEY = constants.PARAMETER_CONSTANT_ACTIVITY_KEY
+    TIMESTAMP_KEY = constants.PARAMETER_CONSTANT_TIMESTAMP_KEY
+    CASE_ID_KEY = constants.PARAMETER_CONSTANT_CASEID_KEY
+    POPULATION_SIZE = "population_size"
+    ELITISM_RATE = "elitism_rate"
+    CROSSOVER_RATE = "crossover_rate"
+    MUTATION_RATE = "mutation_rate"
+    GENERATIONS = "generations"
+    ELITISM_MIN_SAMPLE = "elitism_min_sample"
+    TOURNAMENT_TIMEOUT = "tournament_timeout"
+    LOG_CSV = "log_csv"
+
+
+class Variants(Enum):
+    CLASSIC = classic
+
+
+def apply(
+    log: Union[EventLog, EventStream, pd.DataFrame],
+    variant=Variants.CLASSIC,
+    parameters: Optional[Dict[Any, Any]] = None,
+) -> Tuple[PetriNet, Marking, Marking]:
+    """
+    Discovers a Petri net using the genetic miner.
+
+    Parameters
+    ---------------
+    log
+        Event log / Event stream / Pandas dataframe
+    variant
+        Variant of the algorithm to be used, possible values:
+        - Variants.CLASSIC
+    parameters
+        Variant-specific parameters
+
+    Returns
+    ---------------
+    net
+        Petri net
+    im
+        Initial marking
+    fm
+        Final marking
+    """
+    return exec_utils.get_variant(variant).apply(log, parameters)

pm4py/algo/discovery/genetic/util.py

@@ -0,0 +1,87 @@
+'''
+    PM4Py – A Process Mining Library for Python
+Copyright (C) 2026 Process Intelligence Solutions UG (haftungsbeschränkt)
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU Affero General Public License as
+published by the Free Software Foundation, either version 3 of the
+License, or any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU Affero General Public License for more details.
+
+You should have received a copy of the GNU Affero General Public License
+along with this program.  If not, see this software project's root or
+visit <https://www.gnu.org/licenses/>.
+
+Website: https://processintelligence.solutions
+Contact: info@processintelligence.solutions
+'''
+
+# Author: Maximilian Josef Frank (https://orcid.org/0000-0002-0714-7748)
+
+import random
+import itertools
+
+# typing
+from collections.abc import Iterable
+InputMap = dict[str,list[frozenset]]
+OutputMap = dict[str,list[frozenset]]
+Individual = tuple[InputMap,OutputMap]
+
+class iset(frozenset):
+    "Indexable frozenset printing as set, i.e. without `frozenset(…)`"
+    def __repr__(self):
+        return "{" + repr(sorted(self))[1:-1] + "}"
+
+    @staticmethod
+    def flat(item: Iterable) -> Self:
+        return iset(itertools.chain(*item))
+
+def rand_partition(pool: Iterable) -> list[set]:
+    pool = set(pool)
+    #     also ensures no activity in two partitions
+    #     s. 4. Causal Matrix, Def. 4; https://doi.org/10.1007/11494744_5
+    partition = []
+    while pool:
+        draw = iset(random.sample(
+            tuple(pool),
+            random.randint(1, len(pool))
+        ))
+        partition.append(draw)
+        pool -= draw
+    return partition
+
+def get_src_sink_sets_for_wfnet(I: InputMap, O: OutputMap, T: list[str]) -> tuple[list[str],list[str]]:
+    """Determines input set and output set, which need to be connected by a place to create a WF-net"""
+    def add2graphs(graphs, t, nextT):
+        # find graph
+        graph = next((g for g in graphs if t in g), None)
+        if graph is None:
+            graph = [t]
+            graphs.append(graph)
+        # append plain/grouped T
+        successors = []
+        for S in nextT[t]:
+            if type(S) == str:
+                successors.append(S)
+            else:
+                successors.extend(S)
+        graph += successors
+        # merge if end = start
+        for tn in successors:
+            for g2 in graphs[:]: # [:] = copy
+                if g2[0] == tn and g2 != graph:
+                    graph += g2
+                    graphs.remove(g2)
+                    break
+        return graphs
+    graphsI, graphsO = [], []
+    for t in T:
+        graphsI = add2graphs(graphsI, t, I)
+        graphsO = add2graphs(graphsO, t, O)
+    first = [g[0] for g in graphsO] # ⋃first = reachable via O[∀t]
+    last = [g[0] for g in graphsI] # ⋃first = reachable via I[∀t]
+    return first, last

pm4py/algo/discovery/genetic/variants/__init__.py

@@ -0,0 +1,22 @@
+'''
+    PM4Py – A Process Mining Library for Python
+Copyright (C) 2026 Process Intelligence Solutions UG (haftungsbeschränkt)
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU Affero General Public License as
+published by the Free Software Foundation, either version 3 of the
+License, or any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU Affero General Public License for more details.
+
+You should have received a copy of the GNU Affero General Public License
+along with this program.  If not, see this software project's root or
+visit <https://www.gnu.org/licenses/>.
+
+Website: https://processintelligence.solutions
+Contact: info@processintelligence.solutions
+'''
+from pm4py.algo.discovery.genetic.variants import classic