=Paper= {{Paper |id=Vol-1867/w13 |storemode=property |title=Endowing Business Artifacts with a Normative Coordination Layer |pdfUrl=https://ceur-ws.org/Vol-1867/w13.pdf |volume=Vol-1867 |authors=Matteo Baldoni,Cristina Baroglio,Federico Capuzzimati,Roberto Micalizio |dblpUrl=https://dblp.org/rec/conf/woa/BaldoniBCM17 }} ==Endowing Business Artifacts with a Normative Coordination Layer== https://ceur-ws.org/Vol-1867/w13.pdf

Endowing Business Artifacts with a Normative Coordination Layer

Matteo Baldoni, Cristina Baroglio, Federico Capuzzimati, Roberto Micalizio
Università degli Studi di Torino — Dipartimento di Informatica
c.so Svizzera 185, I-10149 Torino (Italy)
Email: {firstname.lastname}@unito.it

Abstract—We propose to enrich the artifact-centric approach is a fact that business artifacts encapsulate data which are
in two ways. First, by relying on the Agent-Oriented Paradigm, created and manipulated by many processes. For instance an
the tasks acting on artifacts are organized in agents, seen as order is created by a client who interacts with a seller, and
autonomous loci of control, whose execution is goal-driven. is manipulated by the operations that make the transaction
Second, the business artifact model is complemented by a between the two proceed. The client, the seller, more in
normative dimension. Norms are used to represent the data general the processes that operate on a business artifact need
lifecycle in a form that is inspectable and that can be reasoned to agree on who should do what and sometimes even when
upon by agents. Agents can therefore create expectations about to carry out a task of interest – e.g., payment and delivery
the behaviors of others and hence, leveraging on the norms, must both occur otherwise the purchase transaction will not
agents can act on an artifact so as to entice, or oblige, others reach a happy end, and the two tasks are up to different
to act themselves. The paper discusses the advantages and parties in the transaction. So, on the one hand, a business
consequences of this norm-aware enrichment, and outlines a artifact has a lifecycle that describes its evolution from
possible realization based on social commitments. creation to some state where it is considered as archivable.
On the other hand, along the lines of activity theory –
1. Introduction at the basis of A&A – a business artifact should also be
a medium of coordination and interaction. However, this
The artifact-centric approach [5], [10], [8] is recently latter dimension is not supported by state-of-art literature
emerging as a viable solution for specifying and deploying on business artifacts. The proposal we present in this paper
business operations by combining both data and processes aims at overcoming the described lack of business artifacts
as first-class citizens. In particular, the notion of Business with a multiagent systems (MAS) approach. Specifically, we
Artifact, initially proposed by Nigam and Caswell [12], claim that: (1) services by which it is possible to operate on
opened the way to the development of a data-driven ap- business artifacts should be encapsulated and organized into
proach to the modeling of business operations. The data- goal-oriented containers; (2) it is necessary to introduce a
driven approach counterposes a data-centric vision to the normative layer to capture the behaviors that are expected
activity-centric vision, traditionally used when processes of the parties. The paper motivates our research and our
are explicitly modeled in terms of workflows. Artifacts are proposal, and illustrates it with the help of a simple purchase
concrete, identifiable, self-describing chunks of information, example.
the basic building blocks by which business models and
operations are described. They are business-relevant objects 2. Motivations
that are created and evolve as they pass through business op-
erations. A business artifact includes an information model In order to understand how business artifacts are cur-
of the data, and a lifecycle model, the latter capturing the rently specified and used, we briefly introduce the BALSA
key states through which data evolve together with state methodology [6] as a significant representative of the current
transitions. The lifecycle model is used both at runtime approaches to business artifacts. The BALSA methodology
to track the evolution of artifacts, and at design time to specifies a data-centric declarative model of business oper-
distribute tasks. The presence of an explicit lifecycle gives ations, and can be summarized in three steps: 1) identify
business artifacts a semantics that differentiates them from the relevant business artifacts of the problem at hand and
other programming abstractions, like objects, active objects, their lifecycles, 2) develop a detailed specification of the
and artifacts in the sense given to the concept by the A&A services (or tasks) that will cause the evolution of the
meta-model [13]. The lack of autonomy differentiates them artifact lifecycles, 3) define a number of ECA-rules (Event-
from agents. Condition-Action) that create associations between services
The work presented in this paper attacks a limit that and artifacts. ECA-rules are the building blocks to define,
business artifacts show: they do not provide the programmer in a declarative way, processes operating on data.
with any means for designing and modularizing the coordi- BALSA (and similar) is extremely interesting, in par-
nation of those processes which should operate on them. It ticular because it introduces a novel perspective on the
72

modeling of business processes. However, for what concerns that: (1) services should be encapsulated and organized into
coordination in presence of more business processes, the goal-oriented containers; (2) it is necessary to introduce a
methodology simply refers to choreography languages and normative layer. For what concerns (1), the Agent-Oriented
techniques proposed for service-oriented computing, despite Paradigm is a good candidate. In particular, the Agent
the presence of the business artifacts which could themselves and Artifact meta-model (A&A) [13] has already shown
be natural instruments of coordination. how artifacts can be used as environment components that
We deem the absence of an explicit use of the business mediate agents’ interactions. However, artifacts in the A&A
artifacts to the aims of coordination as a significant flaw. In model are radically different from the business artifacts
particular, in inherently destructured settings – like cross- because they do not come with an explicit information
organizational settings–, the involved actors are all peers, model for data, and they do not exhibit data lifecycles.
each of which has its own business goals, and acts in an Thus, this information cannot be exploited at design time,
autonomous way. Each actor does not know and does not nor at runtime, to reason about which actions an agent
care about the possible goals of others. Nevertheless, actors should take. Concerning (2), the normative layer would
generally need to interact to achieve goals they would not be provide an explicit representation of the business artifacts
able to achieve alone. The interaction is a critical dimension lifecycles, and of how coordination is expected to occur.
that need to be explicitly modeled to coordinate the usage of Such a representation would allow agents to reason about the
shared resources. This poses the question of how to scale the use of business artifacts and to create mutual engagements
business artifact model to coordinate autonomous entities. for driving their activities. Indeed, we envisage engagements
We see in the introduction of a coordination model within as encoding causal relations between the actions of an agent
business artifacts the way to achieve this goal, and explain and the goals and actions of another, with a normative
what we mean with a simple example. power that would allow each agent to have expectations on
Let us consider a purchase scenario, involving a mer- the behavior of the others. In the purchase example, it is
chant and a client. We claim that in order to coordinate the easy to see how the introduction of a norm in form of the
interaction between the two agents, it is necessary to add commitment whenever a customer pays, the merchant will
to the plain message exchange (which standard approaches ship the goods, would enhance coordination. The customer
to business processes envisage as the only means of inter- now knows that after service pay, the merchant will be
action), one further abstraction that explicitly represents the pushed to consider the service ship-goods as one of its
engagements each player has towards the other. We also next goals, otherwise it will violate the norm and will be
claim that business artifacts should trace such engagements sanctioned for that. This provides the customer a guarantee
and their evolution, in order to enable an effective agent co- about the achievement of its own goal (or to recoup its
ordination. For example, when offering to sell some goods, losses). An explicit normative layer plays a central role both
the merchant commits to the client to ship the items the at the design time, to verify whether all the engagements
client will pay for. Such a commitment is stored by the can converge towards their satisfaction, and at running time
business artifact involved in the interaction between the two. to monitor the execution of a system and determine the
Because of his awareness of such a commitment, the client, violation of engagements. In this paper we introduce the
having paid for the goods, expects the shipment to occur. If notion of normative business artifacts as a means to extend
this does not happen, the commitment progresses into a state the artifact-centric approach with a normative layer, where
of “violation” and this information, stored in the business engagements and norms are expressed in terms of social
artifact, provides a proof of the merchant’s misbehavior. commitments [14]. The introduction of a normative layer in
From a different perspective, a client is enticed to use the more general setting of business processes is seen as
a business artifact by the merchant’s commitment, which desirable also in [16].
makes explicit the course of interaction the merchant binds
to, and creates a right on the client that such an expected 3. Coordination via Normative Business Arti-
course of action be respected (i.e., my payment will put an facts
obligation on the merchant to ship the bought items or the
merchant will violate the commitment). On the other hand, Business artifacts are, by definition, data-aware: they
the merchant uses commitments inside the business artifact consider data as a first-class primitive that drives the con-
to entice interactions with potential clients – indeed, the struction of process models [5]. Artifacts, however, are not
obligation yielded by a commitment is activated only if a an end in themselves: they are business relevant entities
client pays for some goods. that are created, accessed, and manipulated by different
In the example, the commitments that go along with a services along a business process. We now show how to
business artifact make explicit the behavior the agents are introduce a normative layer so that business artifacts support
expected to stick to. They also have a normative flavour, as coordination.
diverging behaviors will be considered as violations. This Destructured business processes call for a modulariza-
awareness causes agents to take part to an interaction only if tion of the control flow. Agent-oriented programming [7],
they are fine with the commitments. As such, commitments [19] is conceived exactly for handling multiple and con-
provide a standard to define standards of interaction medi- current control flows. Two elements are central in agent-
ated by business artifacts. To realize this vision, we claim oriented programming: the agents and the environment.
73

Agents, as abstractions of processes, possess their own operations, and where regulative norms are used to create
control flow, summarized as the cyclic process in which expectations on the overall evolution of the system (agents
an agent observes the environment (updating its beliefs), behavior and environment evolution).
deliberates which intentions to achieve, plans how to achieve
them, and finally executes the plan [7]. Beliefs concern the 3.1. Environment/Information systems based on
environment. Intentions lead to action [19], meaning that normative business artifacts
if an agent has an intention, then the expectation is that it
will make a reasonable attempt to achieve it. In this sense,
Figure 1 describes the high-level architecture of the
intentions play a central role in the selection and the execu-
kind of system we imagine: (1) involving business artifacts
tion of actions, which represent the innate capabilities agents
and agents (with their goals), and (2) holistically norm-
have to modify their environment. Among others, (business)
aware. Agents interact with each other and with the en-
artifacts (see A&A-meta model [13]) are privileged elements
vironment by creating and modifying data which belong
of an environment. In particular, in contexts where agents
to an information system and that are reified by business
cannot achieve their goals on their own, but need to interact
artifacts. They are goal-driven and capable of coordinating
with other agents to do so, artifacts provide shared resources
with other agents by creating and exploiting commitments,
that agents will use to mediate their interactions.
obligations, permissions, and prohibitions. The conceptual
We claim that business artifacts should be norm-aware
model of the information system is described in terms
in two ways. First, the lifecycle of a business artifact should
of the norms that regulate the evolution of data, that is,
be made explicit by way of norms that specify how data
data lifecycles, capturing how data pass from one state to
evolve. The agents (i.e., the artifact users), will be able
another as a consequence of actions that are performed
to inspect and reason upon them to decide if and how
by some agent. Moreover, business artifacts will include
to operate on an artifact to obtain some result. Second,
all those normative elements that regulate the coordination
agents need to coordinate and regulate their interaction while
of the agents that interact by way of the artifact. All this
using the business artifacts to achieve their goals. Given
information is available to the interacting agents in a form
these two bodies of norms, agents will apply reasoning
that allows agents to reason on it. The agents are aware of
techniques to plan proper coordination that, possibly without
the current state (of the lifecycle) of the data, as well as
violating any norm, will lead to goal achievement. This is
of the obligations, prohibitions, commitments, permissions
possible because norms enable the creation of expectations
put on them, and thus they are aware of the tasks expected
and commitments among agents. Moreover, given an explicit
of them and of their parties. At any time it is possible to
representation of such elements it will be possible to realize
check the execution, identifying pending tasks and who is
systems of accountability to discourage or to detect and
responsible of them, as well as possible violations (e.g. of
explain deviant behavior [4].
obligations or commitments), which may activate procedures
Even though data-awareness and norm-awareness are by
specifically designed to handle the case.
and large orthogonal to BDI [19] notions, it is natural to
think of agents as BDI agents for a seamless integration
of all the aspects of deliberation, including the awareness 4. Building Normative Business Artifacts in
of data and of their lifecycles. For instance, an agent, that JaCaMo+
is involved in handling orders, may conclude that, since it
has to pick up three items in the warehouse, since each such In this section we explain how the normative busi-
item is to be packed, since all packagings are performed by a ness artifact we propose can be implemented by relying
same other agent, and since one of its goals is saving energy, on the 2COMM/JaCaMo+ framework [3]. We refer to an
it is preferable to pick them up altogether, and deliver implementation where the BDI agents are implemented in
them to the other agent only afterwards, instead of picking the Jason agent programming language, and where agents
and delivering one item at a time. Data-awareness here is share artifacts, whose creation and manipulation involves an
awareness that three items of a same kind are requested. explicit creation and manipulation of social commitments
Norm-awareness that items are picked because each of them [14]. Social commitments provide the normative layer and
is part of some order, whose lifecycle says that after being enable the coordination of the goal-driven agents.
picked they will be packed. Again data-awareness allows We exemplify the implementation in the purchase sce-
our agent to know that all parcels are to be made by a same nario. In this scenario each agent has its own goals: the
other agent. merchant has the goal of selling goods, while the customer
Relying on agent-oriented programming is promising has the goal of getting some goods. We show how they can
also because a the agent-based model allows to naturally achieve their goals by using a business artifact as the only
tackle the issue of coordination by introducing the concept means of coordination. To this end we need to present both
of norm [18]. The deliberative cycle of agents is affected by sides of the interaction: the normative business artifact, on
the norms and by the obligations these norms generate as the one side, and how the agents use it, on the other side.
a consequence of the agents’ actions. The limit of current Let us consider first the business artifact. Figure 2 shows
agent-based approaches is that they provide no holistic pro- the business artifact representing the transaction occurring
posal where constitutive norms are used also to specify data between a merchant and a customer. This artifact follows the
74

Figure 1. Environment/Information system based on normative business artifacts.

principles of the business artifact proposed in the BALSA inspected by the agents. Specifically agents will be notified
meta-model: a data model is defined to trace relevant pieces of the changes to the business artifact state which include
of information; namely, the merchant and customer iden- changes occurred to the commitments. Among other events,
tifiers, the item that can be sold by the merchant and the they will be aware of the detachment of commitments of
maximum number of pieces that are available. While this which they are debtors, and of the satisfaction (violation) of
information is provided at the time the business artifact is commitments of which they are creditor.
created, three further pieces of information (namely, quo- commitment ShipGoods m e r c h a n t t o c u s t o m e r
tation, quantity and order) are, instead, the result of the create quote ( quantity , customer )
detach a c c e p t ( q u o t a t i o n , q u a n t i t y , customer )
operations above the artifact performed by the agents using discharge ship ( customer )
it. These operations are captured by the business artifact r e l e a s e r e j e c t ( price , quantity , customer )
lifecyle showed here as an automaton: the customer asks commitment E m i t R e c e i p t m e r c h a n t t o c u s t o m e r
the merchant for a quotation of a given quantity of items it create quote ( quantity , customer )
wants to buy. Once the quotation is provided, the customer detach paid ( customer )
discharge emit ( customer )
either decides to reject or accept the quotation. In the first r e l e a s e r e j e c t ( price , quantity , customer )
case, the business artifact achieves a final state and can be
archived. In the second case, a new order number is created commitment PayForGoods c u s t o m e r t o m e r c h a n t
create accept ( quotation , quantity , customer )
to trace the shipping and the payment of the goods. Note detach s h i p ( customer )
that no ordering between shipping and payment is imposed. discharge paid ( customer )
After the payment, the merchant issues the payment receipt, Listing 1. The normative layer in the purchase scenario
and the business artifact can be archived. In our example, the normative layer is given by the set
As noted above, such a business artifact is not suffi- of commitments in Listing 1. For the sake of readability,
ciently rich to trace the causal relationships. The customer the commitments are expressed following the syntax of
may have an expectation about the behavior of the merchant, the Cupid language [9]. The first commitment, ShipGoods,
raised by its experience in purchasing things, that the pay- is created by the merchant when it executes the quote
ment of an item will be followed by the merchant giving operation. That is, besides giving a value to the quotation
the item but this expectation has no normative power. We, information, the quote operation also commits the merchant
therefore, extend this artifact with a normative layer that, as towards the customer. Such a commitment is discharged
anticipated, is expressed in terms of a set of commitments. when the customer accept the quotation, and discharged
A social commitment C(x, y, s, u) captures that agent x when the merchant ships the bought goods to the customer.
(debtor) commits to agent y (creditor) to bring about the Also the second commitment, EmitReceipt, is created by
consequent condition u when the antecedent condition s the merchant by the execution of the quote operation. In
holds (s and u are conjunctions or disjunctions of events). this case, the commitment is detached when the customer
Only the debtor of a commitment can create it. When pays for the goods, and it is discharged when the merchant
s is true the commitment is detached and turns into an emits the receipt towards the customer. Both these two com-
obligation on the debtor. When u is true the commitment is mitments are released by the customer when it rejects the
satisfied. A detached commitment that is canceled or whose quotation provided by the merchant. The last commitment,
consequent becomes f alse is violated. PayForGoods, is created by the customer when it accepts the
To realize a normative business artifact, thus, it is suffi- merchant’s quotation. The commitment is detached when the
cient to associate each operation on the business artifact merchant has shipped the goods, and discharged when the
with operations (e.g. create, discharge, etc.) on one (or customer pays for them.
more) commitment(s). It follows that a normative business Now, let us briefly review the resulting normative busi-
artifact, besides representing the chunk of information at ness artifact. As before, the merchant advertises, by creating
hand, maintains also the created commitments, that can be the artifact, some item to be sold, and specifies the number
75

Information model:
order
merchant customer item max avail quantity quotation order payed

pay ship

emit

accept quot. order receipt
accepted ship pay processed emitted
request quote
start
bus. art. requested sent order shipped
reject
created quotation quotation

quot. rejected

Figure 2. The business artifact for the purchase scenario.

of available units. An interested customer, by inspecting the the data required for the interaction, and the commitments,
artifact, can now see the commitments that the merchant is together with their states, that are created and manipulated
willing to take towards the customer. That is, the customer along the interaction.
can create expectations about the merchant’s behavior that Let us now discuss the other side of the interaction,
have a normative power. The customer is, thus, enticed to that is, how the agents use a normative business artifact. To
accept the quotation because the presence of the commit- exemplify the agents, we use here the JaCaMo+ framework,
ment, as part of the information provided by the business consisting in the well-known JaCaMo multi-agent platform
artifact, yields that this action will create an obligation on enriched with commitment protocols provided by means of
the merchant to deliver the goods that will make it achieve the 2COMM framework. Below, an excerpt of the merchant
his goal. On the other hand, the customer will see also agent program. In this first plan, the merchant is solicited
the commitments it will take in favor of the merchant, to act by the reception of a requestedQuotation event, that
should it join the interaction. So, if the customer starts comes from a customer through the business artifact. The
an interaction by requesting a quotation for a number of body of the plan consists in the execution of quote, which
units, the merchant will provide such a quotation and, at sends a quotation to the customer and causes the creation of
the same time, it will create a commitment to: 1) ship the merchant’s commitments ShipGoods and EmitReceipt.
the goods, provided that the customer accepts the quotation The second plan captures the detachment of the ShipGoods
(ShipGoods), and 2) emit a receipt upon the payment for commitment. The detach of the commitment is indeed an
the goods (EmitReceipt). event generated by the artifact the merchant is focusing on,
Note how the operations performed by agents on the and it is the consequence of the accept operation performed
business artifact make the commitments progress. For in- by the customer. The body of the plan consists in the ship
stance, the customer’s acceptance of the quotation has sev- operation. Finally, the third plan captures the detachment of
eral effects: 1) on the data side, a new order number is the EmitReceipt commitment, also in this case the body of
created to trace shipping and payment; 2) the customer com- the plan aims at discharging the commitment.
mits towards the merchant to pay for the goods once they This example shows how the two agents, merchant and
are delivered (PayForGoods); 3) commitment ShipGoods is customer, can interact with each other by using a business
detached, and then the merchant is now asked to ship the artifact as an interaction medium. The normative layer as-
goods. As a final comment about the artifact, note how sures that each agent will be willing to discharge its own
the commitments do not impose any ordering about the commitments to avoid their violation. Notably, each agent
payment and shipping. In fact, the customer could pay as achieves the goal it had at the beginning of the interaction;
soon as it accepts the quotation, assured by the existence of namely, the merchant gains money by selling goods, and
commitment ShipGoods that pushes the merchant to actually the customer has the goods it is interested in. These two
ship the goods. goals, however, are not immediately apparent when one
A natural way to implement normative business artifacts considers the business artifact depicted in Figure 2. The
is to rely on the 2COMM framework [3]. In 2COMM two final states, in fact, just denote situations in which the
normative business artifacts are reified as commitment-based business artifact can be archived, but do not specify any goal
protocol artifacts, that are provided by the framework as achievement by the involved agents. One further advantage
Java classes. Business artifact operations are mapped into of adding the normative layer is the possibility of making
protocol actions, whereas the data dimension is captured explicit some of the goals and services agents make available
by the notional social state that protocol artifacts maintain. to others. This piece of knowledge can be used by the agents
Indeed, the social state kept by a protocol artifact traces both in a practical reasoning step (see [17]) to decide whether to
76

1 + r e q u e s t e d Q u o t a t i o n ( Quantity , Customer Id )
2 <− q u o t e ( U n i t P r i c e * Q u a n t i t y , Q u a n t i t y , C u s t o m e r I d ) .
3 + c c ( My Role Id , C u s t o m e r R o l e I d , a c c e p t ( P r i c e , Q u a n t i t y , C u s t o m e r R o l e I d ) ,
4 s h i p ( C u s t o m e r R o l e I d ) , "DETACHED" ) : e n a c t m e n t i d ( My Role Id )
5 <− s h i p ( C u s t o m e r R o l e I d , Q u a n t i t y ) .
6 + c c ( My Role Id , C u s t o m e r R o l e I d , p a i d ( C u s t o m e r R o l e I d ) , e m i t R e c e i p t ( C u s t o m e r R o l e I d ) , "DETACHED" )
7 : e n a c t m e n t i d ( My Role Id )
8 <− e m i t R e c e i p t ( C u s t o m e r R o l e I d ) .

join the artifact. notion of accountability is rapidly gaining importance since,
when more organizations come into play, it is even more
5. Conclusions important to trace back who is responsible for what. First
steps can be found in [4]. Another promising extension is
The presented work is strictly related to the problem of to understand how agents could plan the use of business
interaction in multiagent systems. In these systems, interac- artifacts for reaching their goals. An initial attempt to use
tion is mainly focused on the modeling of communication social commitments in planning has been discussed in [2],
patterns (protocols), which are concerned with the sequence but business artifacts are yet to be considered. Finally, the
of messages that can be exchanged between two commu- standardized lifecycle of commitments can be the key for
nicating agents, but disregard the information conveyed by developing an agent programming methodology, similar to
these messages. Recent approaches such as HAPN [20] and the one discussed in [1]. The idea is to program agents so
BSPL [15] have started to consider also the information that they can properly tackle part of the events that are gener-
dimension. HAPN is formally based on automata where ated in the business artifacts of their interest; specifically, the
nodes represent states of the interaction and transitions be- state transitions that occur to commitments in which they are
tween nodes represent the messages that can be exchanged. involved. To conclude, we mention RAW-SYS [11], which
Transitions have a complex structure since for each message enriches the prescriptive process model with data-awareness.
it is possible to define a guard condition on message sending. Although RAW-SYS looks similar to a (normative) business
A similar approach is BSPL where the information flow artifact, the objectives of the two models are quite different.
is decomposed in a number of “simple protocols”, each RAW-SYS is essentially a framework for verifying business
defining the schema of the messages that can be exchanged processes taking into account both the control- and the
together with their parameters. Parameter are decorated as in data-flows. A normative business artifact, instead, aims at
or out (meaning it is received or emitted). BSPL provides a coordinating autonomous agents.
formal framework in which it is possible to verify properties
such as liveliness and safety of a protocol. Both HAPN Acknowledgements. This work was partially supported by
and BSPL, however, show some weaknesses in properly the Accountable Trustworthy Organizations and Systems
handling information. In HAPN, for instance, guards, that (AThOS) project, funded by Università degli Studi di Torino
enable message sending, may refer to information which is and Compagnia di San Paolo (CSP 2014).
not carried by the message itself, but rather maintained in
an external information system, which is not an integral part References
of the HAPN proposal, and hence the complete verification
of an interaction is not actually achievable. BSPL, on the [1] Matteo Baldoni, Cristina Baroglio, Federico Capuzzimati, and
other hand, assumes a distributed view of information. Each Roberto Micalizio. Empowering agent coordination with social en-
gagement. In AI*IA 2015, Advances in Artificial Intelligence, LNCS
participant has its own knowledge base, and the progression 9336, pages 89–101, 2015.
of the interaction makes the local knowledge bases evolve.
[2] Matteo Baldoni, Cristina Baroglio, Federico Capuzzimati, and
The problem, in this case, is that each participant has just Roberto Micalizio. Exploiting social commitments in programming
a local view of the information lifecycle. Thus, an agent agent interaction. In PRIMA 2015: Principles and Practice of Multi-
cannot create expectations about the behaviors of other Agent Systems, pages 566–574, 2015.
participants as a consequence of the messages it sends. The [3] Matteo Baldoni, Cristina Baroglio, Federico Capuzzimati, and
approach we propose overcomes these limitations. Business Roberto Micalizio. Commitment-based agent interaction in JaCaMo+.
artifacts abstract an information system, and provide the Fundamenta Informaticae (to appear), 2017.
environment in which the agents, which are autonomous loci [4] Matteo Baldoni, Cristina Baroglio, Katherine M. May, Roberto Mi-
of control, interact. Both business artifacts and agents are calizio, and Stefano Tedeschi. Computational accountability. In
Proceedings of the AI*IA Workshop on Deep Understanding and
first-class components. The autonomy and flexibility of the Reasoning, pages 56–62, 2016.
agents are preserved and supported; moreover, it is possible
[5] K. Bhattacharya, N. S. Caswell, S. Kumaran, A. Nigam, and F. Y.
to reason both on the evolution of the business artifacts and Wu. Artifact-centered operational modeling: Lessons from customer
on the interaction. This work can be extended along three engagements. IBM Syst. J., 46(4):703–721, 2007.
main lines of research. First of all, an explicit normative [6] K. Bhattacharya, R. Hull, and J. Su. A data-centric design method-
layer paves the way to formal verification techniques for ology for businessprocesses, pages 503–531. Handbook of Research
cross-organizational business processes. In this respect, the on BusinessProcess Modeling. IGI Publishing, 2009.
77

[7] Michael E. Bratman. What is intention? In P. Cohen, J. Morgan, and
M. Pollack, editors, Intensions in Communication, pages 15–31. MIT
Press, Cambridge, MA, 1990.
[8] Diego Calvanese, Giuseppe De Giacomo, and Marco Montali. Foun-
dations of data-aware process analysis: a database theory perspective.
In Proceedings of the 32nd ACM SIGMOD-SIGACT-SIGART Sympo-
sium on Principles of Database Systems, PODS, pages 1–12. ACM,
2013.
[9] Amit K. Chopra and Munindar P. Singh. Cupid: Commitments in
relational algebra. In Blai Bonet and Sven Koenig, editors, Proceed-
ings of the Twenty-Ninth AAAI Conference on Artificial Intelligence,
January 25-30, 2015, Austin, Texas, USA., pages 2052–2059. AAAI
Press, 2015.
[10] David Cohn and Hull Richard. Business Artifacts: A Data-centric
Approach to Modeling Business Operations and Processes. IEEE
Data Eng. Bull., 32(3):3–9, 2009.
[11] Riccardo De Masellis, Chiara Di Francescomarino, Chiara Ghidini,
Marco Montali, and Sergio Tessaris. Add data into business pro-
cess verification: Bridging the gap between theory and practice.
In Proceedings of the Thirty-First AAAI Conference on Artificial
Intelligence, February 4-9, 2017, San Francisco, California, USA.,
pages 1091–1099, 2017.
[12] A. Nigam and N.S. Caswell. Business artifacts: An approach to
operational specification. IBM Systems Journal, 42(3):428 – 445,
2003.
[13] Andrea Omicini, Alessandro Ricci, and Mirko Viroli. Artifacts in
the A&A meta-model for multi-agent systems. Autonomous Agents
and Multi-Agent Systems, 17(3):432–456, December 2008. Special
Issue on Foundations, Advanced Topics and Industrial Perspectives
of Multi-Agent Systems.
[14] Munindar P. Singh. An ontology for commitments in multiagent
systems. Artif. Intell. Law, 7(1):97–113, 1999.
[15] Munindar P. Singh. Information-driven interaction-oriented program-
ming: BSPL, the blindingly simple protocol language. In 10th Inter-
national Conference on Autonomous Agents and Multiagent Systems
(AAMAS), pages 491–498, 2011.
[16] Munindar P. Singh. NoBPM: Supporting Interaction-Oriented Au-
tomation via Normative Specifications of Processes, 2015. Invited
talk, BPM.
[17] Pankaj R. Telang, Munindar P. Singh, and Neil Yorke-Smith. Relating
Goal and Commitment Semantics. In Post-proc. of ProMAS, volume
7217 of LNCS. Springer, 2011.
[18] Göran Therborn. Back to norms! on the scope and dynamics of norms
and normative action. Current Sociology, 50:863–880, 2002.
[19] Michael J. Wooldridge. Introduction to multiagent systems, 2nd
edition. Wiley, 2009.
[20] Nitin Yadav, Lin Padgham, and Michael Winikoff. A tool for defining
agent protocols in HAPN: (demonstration). In Proceedings of the
2015 International Conference on Autonomous Agents and Multiagent
Systems, AAMAS, pages 1935–1936, 2015.