Glossary

Variables used to reparametrise global parameters and/or initial markings and colourings. A variable may be shared between any global parameter, initial marking and colouring field.

Broader: Scenario parameters

See also: Scenario parameters, Scenario

Bounds over a search parameter. For the search parameter, the bounds are of the form , where both and are numeric.

Example: for search parameter , the bounds can be .

See also: Search parameters, Local search space

A scenario defined relative to another one, its base: scenario B derived from scenario A keeps A's values for some scenario parameters and modifies the values of the rest.

The modified parameters are the ones an experiment varies, while the kept ones hold A's values throughout. A simulation experiment sweeps over the modified parameters' values; an optimisation experiment takes them as its search parameters and searches over them.

Broader: Scenario

See also: Scenario, Scenario parameters, Simulation experiment, Optimisation experiment

A sequence of steps each of which runs the Petrinaut simulator and computes a metric. Petrinaut can perform simulation and optimisation experiments.

See also: Simulation experiment, Optimisation experiment

The product of all local search spaces.

Requires: bounds

Broader: Search space

See also: Local search space, Search parameters

The space defined by a search parameter's bounds.

Example: for search parameter , the bounds induce the local search space that lies in.

Requires: bounds

Broader: Search space

See also: Global search space, Bounds, Search parameters

A function evaluated at every run and time. It takes as input the current state of the Petri net and optionally any scenario parameter.

See also: Optimisation objective, Scenario-dependent metric, Scenario-independent metric

A collection of:

• places and their dynamics
• transitions and their kernel, guard and firing rules
• arcs between them.

See also: Model version, State

A snapshot of the model.

See also: Model, Run

A solution that optimises one or more optimisation objectives.

Broader: Solution

See also: Solution, Optimisation objective, Optimisation

The process of maximising or minimising a function with respect to its arguments.

Broader: Solution

See also: Solution, Optimisation objective

An experiment whose aim is to optimise one or more optimisation objectives with respect to one or more search parameters for a given model version. At each step a new solution is suggested and a scenario is fixed from it - and the optimisation objectives are evaluated by running the Petrinaut simulator one or more times.

As in a simulation experiment, the search parameters are usually the ones a derived scenario modifies, with the base scenario's values held constant throughout the search.

Requires: model version + optimiser settings + scenario + search parameter(s) + optimisation objective(s)

Broader: Experiment

See also: Simulation experiment, Optimisation, Search parameters

Also known as Objective function

A choice of an aggregated metric (a real value/scalar per run) to optimise.

See also: Metric, Optimal solution

A choice of optimisation algorithm together with stopping conditions for terminating an optimisation experiment and a Petrinaut simulation budget per optimisation step.

See also: Optimisation, Optimisation experiment

Also known as Execution, Simulation

A run is one play-through of a scenario, identified by its seed. The total number of runs is .

See also: Time, Trace, State, Scenario

A named set of all scenario parameters and their fixed values — every scenario parameter is given exactly one value, leaving a run of a model version nothing to choose but its seed.

A scenario can be written from scratch or derived from another one.

See also: Scenario parameters, Derived scenario, Run

The collection of global (net) parameters, initial place markings and colourings, and auxiliary variables.

Every scenario parameter can be categorical, discrete (a natural or integer number) or continuous (a real number).

See also: Scenario, Auxiliary variables, Search parameters

A metric that depends directly on scenario parameters — it takes one or more scenario parameters as input.

Broader: Metric

See also: Scenario-independent metric, Scenario parameters

A metric that does not depend on scenario parameters at all.

Broader: Metric

See also: Scenario-dependent metric, Scenario parameters

Also known as Free parameters, Optimised parameters

Parameters that we optimise metrics over. They are a subset of scenario parameters.

Broader: Scenario parameters

See also: Bounds, Search space, Solution

The space that search parameters range over. It is defined on a search parameter basis (local search space) or for all search parameters together (global search space).

Requires: bounds

See also: Bounds, Search parameters, Local search space, Global search space

An experiment whose aim is to understand the impact of scenario parameters on one or more metrics for a given model version. At each step a scenario is fixed and the Petrinaut simulator is run one or more times computing metrics.

The parameters it sweeps over are usually the ones a derived scenario modifies, so that everything the base scenario fixes is held constant across the sweep.

Requires: model version + scenario + metric(s)

Broader: Experiment

See also: Optimisation experiment, Trace, Metric

Also known as Search parameter configuration

A set of search parameters' values — a search parameter configuration. By definition, each search parameter value lies in its corresponding local search space.

See also: Search parameters, Local search space, Optimal solution

Also known as Marking and colouring

At time :

• place-specific: marking (token count per place) and colouring (every token's colour values);
• transition-specific: firing counts and enabled transitions.

See also: Time, Trace, Model

Also known as Frame

The engine's snapshot of a run at one time step. A run is a sequence of frames from up to the horizon. The total number of time steps is .

See also: Run, State, Trace

Also known as Trajectory

The whole sequence of frames for a run — the state over time, from initialisation to termination.

See also: Run, Time, State