<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD v1.0 20120330//EN" "JATS-archivearticle1.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink">
  <front>
    <journal-meta />
    <article-meta>
      <title-group>
        <article-title>A Multi-Perspective Approach for Modeling Workflows</article-title>
      </title-group>
      <contrib-group>
        <aff id="aff0">
          <label>0</label>
          <institution>Zakaria Maamar</institution>
        </aff>
      </contrib-group>
    </article-meta>
  </front>
  <body>
    <sec id="sec-1">
      <title>2.1 Presentation</title>
      <p>The 6-P approach comprises six perspectives (Figure 1): process (WF core), role, resource,
information, place, and time. Each perspective is supported by a meta-model that consists of
several meta-concepts. According to the application domain, eg. FOC, the meta-concepts are
instantiated. We recall that a WF is a set of processes, which themselves are a set of primary
tasks. In what follows, the several perspectives are explained:
- Process perspective: focuses on the tasks that constitute a process and on the execution
chronology of these tasks. The process perspective represents “What to carry out” in a
WF.
- Role perspective: focuses on the organisational actors that participate in the execution of
a process. The role perspective represents “Who is responsible for”in a WF.
- Resource perspective: focuses on the software as well as hardware means that are
committed to achieve the execution of a process. The resource perspective represents
“What to use”in a WF.
- Information perspective: focuses on the data that are manipulated by a process. As with
the resource perspective, the information perspective represents “What to use”in a WF.
- Place perspective: focuses on the distribution of the execution context of a process. The
place perspective represents “Where to carry out”in a WF.
- Time perspective (gathers the other perspectives): focuses on the progress of the
execution chronology of a process, according to a time scale. The time perspective
represents “When to carry out”in a WF.</p>
      <p>locate</p>
      <p>situate</p>
      <sec id="sec-1-1">
        <title>Place</title>
        <p>execute</p>
      </sec>
      <sec id="sec-1-2">
        <title>Role</title>
        <p>assign</p>
      </sec>
      <sec id="sec-1-3">
        <title>Time</title>
      </sec>
      <sec id="sec-1-4">
        <title>Process</title>
        <p>authorise
require
manipulate</p>
      </sec>
      <sec id="sec-1-5">
        <title>Resource</title>
      </sec>
      <sec id="sec-1-6">
        <title>Information</title>
        <p>manage
use
In the 6-P approach, the different perspectives are in interactions (Figure 1, dashed lines). In
what follows, we describe the relationships that support these interactions, except the
interactions that involve time perspective. Nine relationships are identified and summarized as
follows:</p>
        <p>Assign(Process-Role): a process is assigned to actors that have roles in the organisation.
Execute(Process-Place): a process is executed in several places of the organisation.
Thanks to the place concept, the distribution aspect, particularly for a
multiorganisational case, can be illustrated.</p>
        <p>Manipulate(Process-Information): during its execution, a process manipulates several
input as well as output information. In addition, the execution chronology of a process
depends on the values taken by these information.</p>
        <p>Require(Process-Resource): in order to be completed, a process requires several types of
resources.</p>
        <p>Situate(Role-Place): a role is situated in a place of the organisation.</p>
        <p>Manage(Resource-Information): a resource manages several information that are
available in different formats.</p>
        <p>Locate(Resource-Place): a resource is located in a place of the organisation.</p>
        <p>Authorise(Role-Information): a role is characterized by the information it is authorised
to access, whether in read, write, or read/write modes.</p>
        <p>Use(Role-Resource): in order to carry out its responsibilities, a role uses several
resources.</p>
        <p>In Figure 1, time perspective represents a framework that gathers the other five perspectives.
Indeed, time perspective requests periodically these perspectives about their internal states, such
as the tasks in progress from the process perspective, the committed resources from the resource
perspective, the involved actors from the role perspective, etc. This provides a snapshot of the
WF progress. In the 6-P approach, intelligent components, called software agents [3], support the
perspectives’management and interaction. Therefore, six types of agents are needed (Figure 2):
Process-Agent, Role-Agent, Time-Agent, Resource-Agent, Place-Agent, and Information-Agent.
Each agent is associated with a perspective.</p>
        <p>- A Process-Agent, considered as the WF engine, is responsible for process and task
manipulation. Hence, this agent deals with all the aspects that are related to a process, in
terms of initiation, assignation, monitoring, adaptation, and validation.
- A Role-Agent acts on behalf of an actor who fills a role in the organisation. This agent
manages details regarding this actor, in terms of availability, experience, and background.
- An Information-Agent is responsible for information management. Hence, all the
requests of using as well as updating information have to go through this agent.
Moreover, the Information-Agent avoids conflicts on information, by dealing with
concurrent operations.
- A Resource-Agent acts as a proxy on a resource and hence, manages its period of use.</p>
        <p>Furthermore, this agent supports the well functioning of a resource, by requesting its
maintenance when it is needed, for example
- A Place-Agent identifies a place and its content, in terms of roles, resources, and
information. Moreover, this agent returns information on a place’s position, such as
floor number, with respect to other places of the organisation.
- A Time-Agent is in charge of invoking the other perspectives, through their respective
agents. The purpose of this invocation is to know the WF progress.</p>
        <p>Software agent
Process-Agent</p>
        <p>Process</p>
        <p>Role-Agent</p>
        <p>Role</p>
        <p>Information-Agent</p>
        <p>Resource-Agent
Information</p>
        <p>Resource</p>
        <p>Place-Agent</p>
        <p>Place</p>
        <p>Time-Agent</p>
        <p>Time</p>
      </sec>
    </sec>
    <sec id="sec-2">
      <title>2.2 Role perspective</title>
      <p>In the 6-P approach, role perspective identifies the roles that will participate in the WF execution.
Generally, a role identifies a position that is filled by one or several actors. These actors meet this
role’s requirements.</p>
      <p>Figure 3 presents a meta-model of the role perspective. This meta-model is decomposed into two
levels: conceptual level with role, requirement, and service classes, and domain level with actor
and qualification classes. Role class is an abstract description of a position, eg. plan developer in
FOC. In the same figure, a role is in relation with other roles through supervise/supervise-by
relationship. In an organisation, supervision means authority. A role is associated with
responsibilities to achieve, by using different resources (link to resource perspective). In Figure 3,
service class identifies responsibilities. In FOC, responsibilities may include air plans and air
tasking orders. Performing the services of a role means that a WF has been initiated. Actor class
represents either a moral or physical entity, eg. Capt. M. James, who is going to play a role.
Requirement class describes the aspects that are needed to fulfill a role, in terms of expertise,
technical background, and years of experience. Finally, qualification class represents actors’
abilities to meet roles’requirements.</p>
      <p>supervise/
supervise-by</p>
      <sec id="sec-2-1">
        <title>Role</title>
        <p>1
play
+1
Actor
+1
+1 have +1
offer
+1
+1 own +1</p>
      </sec>
      <sec id="sec-2-2">
        <title>Requirement</title>
      </sec>
      <sec id="sec-2-3">
        <title>Service</title>
      </sec>
      <sec id="sec-2-4">
        <title>Qualification</title>
      </sec>
      <sec id="sec-2-5">
        <title>Conceptual level</title>
        <p>When a task is initiated for execution by a process-agent, the next step consists of identifying
potential actors. Each actor is associated with a role-agent. Once these actors are obtained, two
assignation strategies can be used. A first strategy consists of doing an automatic assignation; a
process-agent selects appropriate role-agents. To this end, diverse parameters are taken into
account, such as actor’s availability, actor’s workload, task’s deadline, and task’s priority. A
second strategy consists of doing a manual assignation; a person selects appropriate actors, always
through their role-agents. As with the automatic strategy, the same parameters of selecting actors
could be applied in the manual strategy. In the 6-P approach, the automatic strategy is privileged.
Enhanced with appropriate mechanisms, such as negotiation, and taking into account the
characteristics of the environment in which they evolve, process-agents could be able to decide to
whom assign tasks.</p>
      </sec>
    </sec>
    <sec id="sec-3">
      <title>2.3 Place perspective</title>
      <p>In the 6-P approach, place perspective presents how the execution of a WF is distributed
according to the places of the involved organisations. Figure 4 presents a meta-model of the
place perspective. In this figure, a place is considered as a geographical position that can be
whether an office, a building, or a meeting room (additional types of places could be considered,
if needed). Interesting to note that place perspective could be used to illustrate organisational
dependencies that concern the decision distribution.</p>
      <p>Place</p>
      <p>Office Meeting room Building</p>
      <p>Figure 4 Meta-model of place perspective</p>
    </sec>
    <sec id="sec-4">
      <title>2.4 Application to FOC</title>
      <p>In what follows, the application of the 6-P approach to FOC has been simplified for the purpose
of explanation (Figure 5). The WF that we intend to set up, and hence to model, is called battle
management. It consists of the following processes: assessment of situation and generation of
options. Each process takes place in a specific building; assessment of situation in buildingA and
generation of options in buildingB. Assessment of situation process consists of two tasks:
evaluate events and compile resources. Generation of options process consists of three tasks:
identify options, evaluate resources, and select an option. BuildingA and buildingB contain
respectively two offices and one meeting room. In what follows, we focus on evaluate events
task. A soldier who plays the monitor role carries out this task. He uses sensor application –
version 1.3- that runs on top of a PC station. This application receives formatted messages from
remote sensors, parses these messages, and finally updates appropriate data, among them event
type and event priority. According to time perspective, the state of battle management WF at T1
time is as follows:
- Involved processes/tasks: assessment of situation/evaluate events.
- Involved roles: monitor.
- Involved resources: PC station, sensor-application -version 1.3-.
- Involved places: Office1.</p>
      <p>- Involved information: event_type, event_priority.</p>
      <p>Once evaluate events task is completed, compile resources task is launched. Another soldier who
is in office1 carries out this task. However, to carry out this task successfully, this soldier has to
move for security reasons to office2, in order to use personnel application that accesses personnel
database. In BuildingB, select an option task of generation of options process is handled by the
commander and his executive element at T3 time. This task takes place in meeting-room1.</p>
      <p>Workflow</p>
      <p>Battle
management</p>
      <p>Process
Sensors</p>
      <p>Pers.</p>
      <p>DB</p>
      <p>Assessment
of situation</p>
      <p>Task
evaluate
events
Monitor
PC, sen.-app.</p>
      <p>Event.type
Event.priority</p>
      <p>Office1
compile
resources
Monitor</p>
      <sec id="sec-4-1">
        <title>Office1</title>
        <p>PC
Pers.-app.
Pers.name</p>
      </sec>
      <sec id="sec-4-2">
        <title>Time1</title>
      </sec>
      <sec id="sec-4-3">
        <title>Office2</title>
      </sec>
      <sec id="sec-4-4">
        <title>Time2</title>
        <p>Generation
of options
Task
select
an option
Commander
Exec. staff</p>
      </sec>
      <sec id="sec-4-5">
        <title>Meeting-room1</title>
      </sec>
      <sec id="sec-4-6">
        <title>Time3</title>
        <p>BuildingA BuildingB
Figure 5 Application of 6-P approach to FOC
Time</p>
      </sec>
    </sec>
    <sec id="sec-5">
      <title>5. Summary and Future Research</title>
      <p>In this position paper, we outlined the 6-P approach for modeling WFs that relies on six
perspectives, respectively called process, role, resource, information, place, and finally time.
Software agents that assess their environments before committing themselves manage these
perspectives. The approach is applied to model FOC’s processes.</p>
      <p>Furthermore, the proposed approach could be used in other application domains, in which
distribution and adaptability constraints should be dealt with. For instance, the place perspective
supports distribution while role perspective and software agents support adaptability.
Currently, the 6-P approach has been used in the context of one organisation. Further work
needs to be done in multi-organisational contexts. The following perspectives could be subjects
of research:</p>
      <p>Role perspective: in an organisation, roles interact together; they invoke their
respective services. Such invocations have to take into account authority as well as
cooperation relationships that could exist between roles. Interesting to analyse
situations in which roles belong to different organisations. Normally, agreements
should dictate how roles should behave, for instance by specifying who is authorised
to answer whom? To this end, a social network, linking the role perspectives, should
be built.</p>
      <p>Information perspective: in order to handle process interactions in a
multiorganisational context, representation formalism, such as XML, of the data to be
exchanged should be used. Furthermore, semantic conflicts because of information
heterogeneity should be dealt with. To this end, a semantic network, linking the
information perspectives, should be built.</p>
    </sec>
  </body>
  <back>
    <ref-list>
      <ref id="ref1">
        <mixed-citation>
          1
          <string-name>
            <given-names>C.</given-names>
            <surname>Hanachi</surname>
          </string-name>
          and
          <string-name>
            <given-names>Z.</given-names>
            <surname>Maamar</surname>
          </string-name>
          . “
          <article-title>Database Agentification for Flexible Workflows”</article-title>
          ,
          <source>in The 3rd. World Multiconference on Systemics, Cybernetics and Informatics and the 5th. Int. Conference on Information Systems Analysis and Synthesis, session on Agent-Based Collaborative Design and Manufacturing</source>
          .
          <source>Florida</source>
          <year>1999</year>
          .
          <string-name>
            <given-names>J.</given-names>
            <surname>Sutherland</surname>
          </string-name>
          and
          <string-name>
            <given-names>S.</given-names>
            <surname>Alpert</surname>
          </string-name>
          . “
          <article-title>Big Workflow” for Enterprise Applications</article-title>
          , in Patel, D.,
          <string-name>
            <surname>Sutherland</surname>
            ,
            <given-names>J.</given-names>
          </string-name>
          &amp;
          <string-name>
            <surname>Miller</surname>
            ,
            <given-names>J</given-names>
          </string-name>
          . (eds.),
          <source>Business Object Design and Implementation III: OOPSLA '99 Workshop Proceedings</source>
          . Springer-Verlag. Denver
          <year>1999</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref2">
        <mixed-citation>
          <string-name>
            <given-names>N.</given-names>
            <surname>Jennings</surname>
          </string-name>
          ,
          <string-name>
            <given-names>K.</given-names>
            <surname>Sycara</surname>
          </string-name>
          , and
          <string-name>
            <given-names>M.</given-names>
            <surname>Wooldridge</surname>
          </string-name>
          .
          <article-title>A roadmap of agent research and development</article-title>
          .
          <source>Autonomous Agents and Multi-Agent Systems</source>
          ,
          <volume>1</volume>
          (
          <issue>1</issue>
          ):
          <fpage>7</fpage>
          -
          <lpage>38</lpage>
          ,
          <year>1998</year>
          .
        </mixed-citation>
      </ref>
    </ref-list>
  </back>
</article>