MASCEM's Ontologies
MASCEM Public Ontologies
The necessity for the integration of different models and platforms brings out the need for communication capabilities that allow entities of different environments (such as software agents) to be able to understand each other and cooperate towards a common goal. The use of ontologies grant the communication capabilities needed by representing concepts and relationships, defining a common “language” that can be understood by software agents, allowing them to coexist and collaborate.
The use of ontologies to support players’ communications, providing means for an easier cooperation with heterogeneous agent societies, complement the simulation capabilities of MASCEM. The concepts and their relationships are represented in OWL and can be used and extended by each different simulation platform, in a way to integrate efforts and different perspectives.
The use of languages that can be understood by different systems facilitates the connection and cooperation between them, enabling simulators, such as MASCEM, to integrate several different electricity markets' (EM) models and power system approaches that allow a broader study capability in this field. Using such common communication language, agents from heterogeneous systems are able to participate in simulations performed by other systems, and use computational models that until now were only available to entities of the same system.
Ontologies for the interoperability of EM multi-agent simulation platforms
It is proposed the use of ontologies for the interoperability of EM multi-agent simulation platforms, which can be extended in a way to enable full interoperability between those systems, as is the example of the ontologies developed for our multi-agent based simulators. Figure 1 illustrates the MASCEM's public ontologies.
Figure 1 - Ontologies for the interoperability of EM multi-agent simulators
The Electricity Markets Ontology is the base ontology imported by all the remaining ontologies. Concepts present in this ontology must be abstract enough to be reused by other ontologies and further extended in each of the EM ontologies, namely the EPEX, MIBEL and Nord Pool ontologies.
The ontologies are formulated in OWL DL and its representation is in RDF/XML. They are publicly available so that they can be used by third-party developers who wish to integrate their agent-based simulators with MASCEM, taking advantage of its simulation capabilities and market models; or with AiD-EM which provides decision support in the scope of the EM. On the other hand, the ontologies may also be reused and extended for the development of new multi-agent simulation tools in the context of the electricity wholesale markets.
Electricity Markets Ontology
The Electricity Markets Ontology (EMO) meets abstract concepts and axioms to the main existing EM. This ontology aims to be as inclusive as possible so it can be extended and reused in the development of (lower level) market-specific ontologies as the MIBEL, EPEX, Nord Pool or any other electricity market ontology. EMO is the base ontology from which all the remaining ontologies developed in the scope of this work extend. This is also the ontology to extend when it is intended to develop the definition of a new market ontology, such as GME (the Italian electricity market) to be possibly included in MASCEM or in other electricity market simulator. Figure 2 illustrates the EMO's concepts and the relations between them.
Figure 2 - Electricity Markets Ontology
Analysing Figure 2 it is possible to see the object properties represented in blue and the data properties defined within each class with the respective data types. The orange relations represent the inferred object properties, which are inverse properties of the ones defined in blue in the opposite direction. It should be noticed that three object properties defined in this ontology are not present in the diagram, namely: hasBilateralContract, placedInPeriod and placedInSinglePeriod. These are important properties that are introduced in EMO to be reused after by the ontologies defined by each electricity market’s domain.
MIBEL Ontology
The MIBEL Ontology (MBL) is related with the Iberian EM. Since it is constantly evolving it is important to keep the ontology as flexible as possible. MBL imports EMO, extending some of it's concepts, and including some new ones as the Complex Conditions.
The MBL ontology is illustrated in Figure 3 where are represented its classes, object and data properties; and the existing relations between them and between them the EMO’s concepts and properties.
Figure 3 - MIBEL Ontology
The yellow rectangles represent the concepts imported from EMO. In blue are the object properties, and if imported from EMO they use the prefix “EMO:”. In orange is represented the inferred object property incorporatesComplexCondition which is an inverse property of incorporatedByPlayer. The relations between the remaining EMO concepts where left out in order to simplify the reading of the diagrams.
EPEX Ontology
Due to the great similarity between the EPEX and Nord Pool markets, both ontologies where developed simultaneously. However it has been decided to develop both separately due to the fast evolution of the markets, which could lead to follow distinct paths one day.
The EPEX Ontology (EPX) imports EMO, extending its concepts and including some new data properties. The classes, the object properties and data properties defined in EPX are shown in Figure 4.
Figure 4 - EPEX Ontology
EMO’s classes are presented in yellow, while its data and object properties have the prefix “EMO:”. As it can be seen, the EMO:Area is redefined in EPX, including the new four data properties of EPX.
Nord Pool Ontology
The Nord Pool Ontology (NPO) is very similar to the EPX, as mentioned above. The main difference is related to the inclusion of the flexible hourly orders. Figure 5 exposes the classes, object and data properties of NPO.
Figure 5 - Nord Pool Ontology
In yellow are illustrated the EMO’s concepts, and the prefix “EMO:” identifies EMO’s object and data properties. In blue are identified the object properties of both EMO and NPO. In NPO the EMO:Area includes the three new added data properties, besides the ones included in EPX. Another point to take into consideration is the FlexibleHourlyOrder which can only be placed in one single session.
Call For Proposal Ontology
After the development of each market's ontology (i.e. MIBEL, EPEX and Nord Pool), and taking advantage of the knowledge already acquired, it has been started the development of the Call For Proposal Ontology (CFP). It imports EMO and extends it by including two new concepts and an object property. Figure 6 demonstrates the relations between concepts of the CFP and EMO ontologies.
Figure 6 - Call For Proposal Ontology
As expressed before, the classes imported from the EMO are in yellow, while the object properties are represented in blue. In order to use this ontology, it should be noted that the CallForProposal class is used by the market operators agents to inform players agents that a given market is about to begin and they should submit their proposals if willing to participate. In response to the call for proposal (CfP) each market player interested in participate in the market sends a Proposal to the market operator with the respective bids and complex offers, if desired.
Electricity Markets Results Ontology
The Electricity Markets Results Ontology (EMR) ontology has been developed considering the output data each market provides its players. Given the simplicity and similarity of the several EM results, it has been decided to gather this knowledge in a single ontology. Figure 7 illustrates the concepts and properties of EMR and its relation to the EMO ontology.
Figure 7 - Electricity Markets Results Ontology
Observing Figure 7, it is visible in yellow the objects imported from EMO, while the object properties are represented in blue. The data properties are identified within each class, and the prefix “EMO:” identifies EMO’s imports. This ontology is used by the market operators to inform the players about their results and outcomes in the market.
AiD-EM Ontology
The AiD-EM Ontology (ADM) has been developed aiming at providing AiD-EM’s interoperability with any electricity market player of any agent based simulation platform. Our goal is, not only provide MASCEM’s players with decision support, but also to make it available for any agent intending to participate in EM simulations.
The ADM ontology also imports EMO concepts and extends it by adding new classes, object and data properties; and relating them with EMO’s players. The relations between concepts of ADM are illustrated in Figure 8.
Figure 8 - AiD-EM Ontology
Following the same nomenclature, the concepts imported from EMO are represented in yellow, in blue the object properties, and the data properties are included in the related classes. The “EMO:” prefix identifies the classes, object or data properties imported from EMO.
As it is possible to observe, AiD-EM includes a large variety of strategies to provide EM players with decision support. There are three main types of strategies from which all the remaining extend, namely MarketStrategy, RLAStrategy and PortfolioStrategy; where each refers to a distinct area and can be used by more than one Tool. The GameTheory market strategy is the only, so far, that makes use of a ScenarioAnalysisMethod.
When a Player requests for AiD-EM’s support, he must choose, from the available strategies, the one(s) to use. ADM has been designed for communication purposes only, between AiD-EM’s agents and EM players requesting for support.