Modeling Collaborative Task Execution In Social Networks2

Modeling Collaborative Task Execution In Social Networks A B Sagar (email:[email protected]) Dept of Computers and Information Sciences, Hyderabad Central University, Hyderabad, India

ABSTRACT A Social Network Group (SNG) in the context of this paper is a group of people who come together to form a small scale business. The relationship between them is business. This paper provides a formal framework for SNGs and also models a rudimentary framework for collaborations among SNGs. The chief functionality of SNGs is to execute tasks, which forms the basic reason for collaborations. So, in the process of developing a framework for collaborations, we first formalize a task and then proceed to develop framework for collaborations. Collaboration is seen as a joint effort of multiple individuals or SNGs to accomplish a task or project. This is an extremely useful feature which helps in achieving goals which are otherwise not feasible for a single individual or SNG. This paper models a rudimentary framework for collaborations and also proposes a metric of collaboration.

keywords: SNG, Social Network Group, Collaboration, collaborativeness, formalization

task

formalization,

collaborations

General Terms: Design Category: K.4.1

1. INTRODUCTION A Social Network Group [SNG] is a group of 10 to 20 people who come together on a business relationship to form a small scale business. The people [hereafter referred to as ‘members’] pool their skills for their business growth and personal development. SNGs can be used to provide the means to distribute nation’s wealth and resources into the society by having government grant projects to them. SNGs also provide several other advantages such as helping members manage cash flow deficits leading to improvement in quality and productivity of their only capital/resource i.e. human capital/resource; helping members avoid money lenders, especially to meet food and health emergencies; helping members invest in asset

Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. CUBE 2012, September 3–5, 2012, Pune, Maharashtra, India. Copyright 2012 ACM 978-1-4503-1185-4/12/09…$10.00.

creation, diversify their occupations, and improve their risk-bearing capacities; promoting leadership qualities among their members; fostering women, even from conservative communities and regions, to interact with outsiders, particularly officials, including men; and also establishing the linkage between banks and marginalized citizens, which is great step for women empowerment. SNGs have to improve the way they develop products. Since their finances and technical capabilities are limited, a collaborated product development is a good solution for SNGs to develop their businesses. SNGs must deal with the need of ever increasing capability and complexity of product lines by incorporating new techniques and technology with constraints of limited time and budget. So more research is indeed essential to develop Collaborative task execution solutions and configuration of processes, people, tools, and structural arrangements for SNGs to achieve the CPD goals. The basic idea is that SNG obtains small projects from public or private sector. Each project is comprised of several tasks. Some tasks that they receive are feasible and some are infeasible. When SNGs receive infeasible tasks, i.e. tasks which they cannot complete with their existing resources [resources mean skills, materials, machinery, members, etc], they need to collaborate with other SNGs who have the required resources. Collaboration is a joint effort of multiple SNGs to accomplish a task or project. Collaboration has more credibility, influence, and ability to accomplish objectives than a single entity. Collaboration with other SNGs provides several other benefits such as lowering production costs, access to less expensive labor, increased creativity and innovation, better products/services, improved revenue opportunity, shrinking distances and time, etc. Thus, collaborations are very beneficial to SNGs. But there are several issues such as communication needs, identity management, performance, mutual benefits, costs, etc. involved in collaborations. Collaboration is something more than mere coordination or cooperation. Understanding the difference between the words coordination, cooperation and collaboration is challenging because they have often been used interchangeably. All these three words include, “working together” in their primary definition. However, each word differs slightly on who is working together, and what they might be working on. Merriam-Webster Online (2009) further defined these words as follows: ∗

Coordination: the harmonious functioning of parts for effective results.

∗ ∗

Cooperation: to act together or in compliance for mutual benefit Collaboration: to work together jointly, especially in an intellectual endeavor;

Collaboration involves identifying the right SNG with most relevant skills, personalities, knowledge, work-styles, and ethical values, ensuring they share commitment to the collaboration task at hand, and offering them environment, tools, knowledge, training, process and facilitation to guarantee they work together efficiently.

2. RELATED WORK Sousa et al. [1] developed a project for social networking system for educational professionals, this paper considered the kind of technologies could be used to create a web application that provided, type of interaction, behavior needed, requirements, technologies and the system implementation. A report [2] also has been submitted which is examining three specific types of collaborative behavior and assessing their impacts on knowledge creation. Various types of social networks, such as those related to professional collaboration [3-5], Internet dating [6], and opinion formation among people have been studied. Social networks involve Financial, Cultural, Educational, Families, Relations and so on. Jihye Song et al. [6] proposed a fundamental framework of discovering and analyzing a workflow-based social network formed through workflow-based organizational business operations. A little more precisely speaking, the framework formalizes a series of theoretical steps from discovering a workflow-based social network to analyzing the discovered social network. The present paper proposes a framework for task execution collaboration in SNGs whose working environment can change dynamically. It studies various communication issues regarding the search of a collaborator and thereof. This lays a foundation for further work in this area of research.

3. ORGANIZATION OF PAPER This paper is organized as follows. Section 1 talks about what an SNG is and what are its basic properties and its usefulness and the importance of collaborations. Section 2 gives the related work done so far in this field. Section 4 talks about the basic idea of how SNGs can be formed using the Publish Subscrible Model. Section 5 talks about formalizing the collaborations, viewing tasks with respect to their states, factors affecting the collaborations, etc. Section 6 is the conclusion of the paper.

4. FORMATION OF SNGS 3.1 PUBLISH-SUBSCRIBE MODEL The Publish-Subscribe Model is the process of collaboration in SNGs. It executes at every SNG. The model considers two basic methods for finding a collaborator: (1) Advertising collaboration specifications and then waiting for collaborators to subscribe (2) Subscribing to other SNGs in the network and finding out suitable collaborators.

Collaboration specification contains resources requirements, collaboration goals, collaboration type, collaboration duration, etc. Collaboration goals are center constructs of the collaboration specification and are fundamental for every collaboration. Resources requirements imply capabilities and process specifications. Capability specification gives the details of the capabilities that a prospective collaborator should possess and process specification details how the task is to be executed i.e. plan of execution of the task. The overlapping between the resources requirement of the advertising SNG, and the excess resources available with another SNG, is the initial factor for the latter SNG to subscribe as a prospective collaborator to the former SNG. Inviting SNGs for collaboration is done by propagation of a collaboration specifications advertisement called collab_advt. The format of the collab_advt is given in figure below. Two basic functions are required for establishing collaboration: SEND function to send a collab_advt or response, and RECEIVE function to receive a collab_advt or response. Every SNG will have these SEND and RECEIVE functions. For example, a SNG ‘A’ sends a collab_advt by SEND(X), another SNG ‘B’ receives it by RECEIVE(X). After analyzing the collab_advt X, ‘B’ sends the response using SEND(X-Response) and ‘A’ receives the response using RECEIVE(X-Response). The SEND/RECEIVE pair are enough for collaboration message passing. Two more functions are also used, STOP and ACK. The STOP function stops advertising. It marks the receipt of enough collaborators’ responses. ACK function is for acknowledging the receipt of response from a responding SNG. Collaboration Process 1. Publish the collab_advt 2. Subscriber Receives collab_advt 3. Subscriber Analyzes the collab_advt 4. Subscriber Responds with accept/reject token 5. If collaborator is found, start collaboration Constructs of the Communication Language Construct & Description Create_Collab_advt - SHG creates an advertisement of collaboration over a task Broadcast - Send the collab_advt over the SHG network Available - A SHG notifies other SHGs about its availability for collaboration Send - The sender sends the collab_advt or response to a specific SHG Receive - The receiver receives the collab_advt or response Response - The receiver responds to the collab_advt Establish_Collab - The sender and receiver agree for collaboration Signal - The sender signals the receiver that it can begin the execution of the task Notify - The receiver or sender notify each other of some information Query - The receiver queries the sender about some information it does not know about the task Inform The sender sends some information to the receiver CannotDoTask - The receiver refuses to perform the task stating the reason for the refusal Alert - The sender alerts the receiver about some event the receiver needs to react to.

Subscribe - The sender requests the receiver to inform the sender whenever The status of the receiver or task changes Unsubscribe - The sender informs the receiver to cancel a previously established subscription

SEARCH FOR COLLABORATOR: Functionally, each SNG is a pool of resources. What makes a SNG a collaborator is the connectedness of the SNG to the task’s need. Hence, finding a suitable collaborator involves finding SNG [or SNGs] with most suitable resources for the task’s need and located at the shortest distance. The problem of finding suitable collaborator(s) now becomes a goal reaching problem through shortest distance. Given a task (infeasible), we need to design a scale-effective cluster of SNGs for solving the problem. We have to search through the entire SNG network and identify the right ‘combination’ of collaborators who can optimally execute the task. This search process can be optimized by limiting the search to a smaller subspace of ‘good’ SNGs. This subspace is formed by moving/promoting well performing SNGs into the subspace. After the best collaborators are chosen, task execution begins based on the task execution plan.

COLLABORATION RATING: Collaboration rating is the penultimate process in the Publish Subscribe Model. When the SNG advertises for collaborators, it receives acceptance from some SNGs that are interested in collaborations, then the SNG needs to make a decision and choose the best collaborator. This rating aids in choosing the collaborators. To create this rating for each SNG, the Publish Subscribe model has this ‘collaboration rating’ step. That is, at the end of each collaboration, the SNG and collaborator rate each other. Thus, every SNG that has participated in a collaboration will obtain a rating. Though several criteria are considered for rating the collaboration, satisfaction degree and trust degree are considered important. SNGs with good collaboration ratings form a subgroup of prioritized collaborators. Usually the best collaborator resides in this subgroup of SNGs. Trust degree can be obtained by computing the ratio of how many tasks a SNG received and how many it completed successfully. A certainty value ‘W’ is derived to compute the satisfaction degree. To compute satisfaction degree of a SNG over a task, multiply the certainty value with a variable factor λ | λ ∈ (0,1). SatDegree(SNG_ID, T ask_Id) = λ ∗ W If the collaborating SNG is satisfied with the collaboration quality, λ is positive, if not satisfied, then λ is zero. The certainty value ‘W’ is computed from several characteristics of the collaborator and they are given in the able below. ‘W’ Computation Table Characteristics of the collaborator - Value Commitment of collaborator towards the task execution - 4 Open-mindedness of the collaborator - 3 Timeliness of the collaborator in sending information - 5 Willingness to continue even during difficult situations- 6 Perceptive capability of the collaborator - 3 Skilfulness at giving/ receiving even negative responses - 4 Self-managing and requires “less supervision” - 3 Thinking differently and bringing new perspectives -5 Comfortability with ambiguity - 4 Adaptiveness to changes/new requirements - 5

Significance of stake in the outcome of collaboration - 4 Ability of the collaborator in improvising performance - 3 Sense of urgency about collaboration - 5 Encouragement of equal participation of all members- 3 Optimism and being on the upbeat - 4 Pragmatism and knowing how ideas get executed - 5 Collaborator has lots of ideas - 3 Collaborator thinks strategically - 5 Collaborator has good knowledge required for the task - 4 Collaborator has good connections and networks - 3 Collaborator has skill to make the group reach consensus-5 Collaborator is well organized - 3 Collaborator is felt trustworthy - 4 Collaborator has experience being a collaborator - 3 Collaborator is gregarious and dynamic - 4 Total = 100

5.

FORMALIZING COLLABORATIONS

Collaborations occur over tasks. So, before formalizing the collaborations, let us formalize the tasks. Tasks are business functions that are executed by all SNGs. Task execution is the basic functionality of SNGs and they are also the basis for collaborations. Once a task was given to a SNG, it is the responsibility of that SNG to execute it and return the result to the issuer. A task, in the present model, is an entity with several attributes. Every task, besides others, has the following key attributes: Task_Id, Parent_Task, IsPartitionable, Risks, Task_Type, Creation_Date, Excluded_Executors, Task_Stakeholders, Task_Status, Skills_Needed, Skills_Levels_Needed, Task_Infrastructure, Task_Priority, Interdependencies, Notification_Recipients, Task_Supervisor, Responsibilities, Task_Initiator, Potential_Executors, Cost_ of_Task, Start_By_Date, Complete_By_Date, etc. Each task [T] has a ‘start state’ and a ‘final state’. Before beginning of execution of a task, the task is at ‘start state’[S]. During execution, the task moves towards completion, and finally reaches the final state called the goal state [G]. The distance between the task’s start state and the goal state is D and it always sums up tp 100. The reason for D to sum up to a 100 is necessary because, we cannot estimate the progress made in task execution based on subtasks since each subtask may have a different complexity. If a task is split into 10 subtasks out of which nine have very less complexity and take only nine days to complete nine of them. But the tenth subtask may itself take ten days. So, when we say that nine subtasks were completed, we were giving a wrong idea that the task is almost complete. Hence we refer the task execution progress in terms of distances travelled towards the goal. The subdistances d1 , d2 , d3 , ... dn are the distances that a task moves towards goal G with the completion of each of its subtasks. Hence, for n subtasks, total distance D is:

Fig 4: Task- Goal- Distance

From Fig: 4, a simplified mathematical representation of a task would be: Task, T = (li , di , tm , n) where li is the initial location of the task, di is the destination of the task, m is the total subtasks and n is the number of collaborators involved. If a task has no subtasks or there exists a constraint that it should not be split into subtasks, then t = 0. And, if a task does not have any collaborators, then n = 0. A task is called a complex task if t > 1 and n > 1 . If the task has several subtasks, then with the completion of each subtask it moves a distance dm . The distance moved after k subtasks is Σdk − li , and remaining distance is obtained from the difference between dk and Σdm . When Σdm = D, the task execution is finished. If complexityi is the complexity involved in executing the task ti , that is, complexityi is the complexity to move the task a subdistance of di , then total complexity of the task, ComplexityT is the summation of complexities of all subtasks. 4.1 Collaborated Distance: A complex task has subtasks and collaborators share the the subtasks. The total distance moved by collaborators is called collaborated distance. The collaborated subtasks may be executed sequentially or nonsequentially. 4.2 Task-Collaboration Graph: This graph gives the detail of the subtasks and the corresponding collaborations. This also details the concurrent subtasks. The subtasks that are enclosed between the vertical lines indicate concurrent subtasks. The weighed edges connecting the subtasks denote the subdistances i.e. the distance a task moves towards the destination when the task is completed. When multiple collaborators execute a subtask, then they all have arrows pointing them. Finally the task reaches the goal state.

4.4 PROCESS FLOW IN A SNG SNG process flow consists of four main parts namely Collaborations Monitor & Communicator, Collaborations Processor, ExTsk and Resources Repository. Collaborations Monitor & Communicator : It is the SNG’s interface for collaborations. It receives collaboration requests and forwards them to Collaborations Processor. And also communicates with other SNGs regarding collaborations. Collaborations Processor: All collaboration requests forwarded from the Collaboration Monitor are analyzed here. This module will communicate with ExTsk and Resources Repository and analyzes the future resources usage. If the SNG has excess resources than needed for the execution of present tasks, then it will indicate the Collaboration Monitor that it is available for collaborations. ExTsk : This is the module logging the status of the running tasks. Resources Repository: All the information regarding SNG’s resources are maintained here. It holds information of the resources currently being used, available resources, and resources in the network that are available through collaborations. When SNGs receive tasks of such size and complexity that no single member or SNG itself is capable of effectively executing them, then they resort to collaborations. A collaboration can be defined as an initiative taken by a SNG to move the task from its start state (S) to (or towards ) the goal state (G) by joint efforts with other SNGs. A ollaborator is the SNG which moves [or helps in moving] the task [or subtask] towards the goal (G). A subdistance di corresponds to a subtask ti [from Fig. 4]. Since the subdistances vary in length with each other, it is more appropriate to measure the progress of the task not by the number of completed subtasks, but by the total distance it moved towards the goal. With addition of collaborators, the goal can be reached more quickly. That is, with the addition of collaborators the total distance D is covered quickly. A SNGi might collaborate with one or more SNGs for executing a task T. If SNGi collaborates with the SNGs SNGj , SNGk , SNGl , SNGm , ...SNG n for executing a task T, then the collaboration is given as :

Fig 5: Task Collaboration Graph 5.3 Task Classification: Task classification helps in defining the complexity of the task. A task is classified as independent if, for its execution and completion, it does not need any collaborators. If the task needs the collaborators then it is a collaborative task. And, if the task is completely infeasible for the SNG, and needs to be transferred to some other SNG, then it is an assignable task. If the collaboration is needed for a short span of time, then it is ‘partial collaboration’, and if the collaboration was needed throughout the task, then it is ‘complete collaboration’. If the collaborative tasks need both parties [ i.e.both the task owner and the collaborator ] to involve in the execution, then it is ‘synchronous’, else if part of the task is executed independently by the collaborator, then it is asynchronous. The spatial or geographic dimension is that the collaborators are either in the same place (co-located) or in different places (remote).

collab(SNG i , SNG {j,k,l,m,n} , T )

Fig 6: Process Flow in a SNG

4.5 FACTORS IMPACTING COLLABORATION n is collaborated tasks; z is total tasks Balance in the collaboration is the collaborators’ view about the fairness of mechanisms, investments, etc in addition to sharing risks and benefits. The relative views of the collaborators about the costs to benefits ratio is the key variable. One collaborator’s view of balance consists of the collaborator’s estimation of equity in resources, the time, and cognitive resources allocated to the partnership. The better the balance, the stronger the collaboration. collaboration_balance(SNGi, SNGj) ⇒ (count(SNGi.tasks Count) ≡ count(SNGj.tasksCount)) ∧ (cost(SNGi.tasks) ≡ cost(SNGj.tasks)) ∧ (benefit(SNGi.tasks) ≡ benefit ( SNGj. tasks)) Fig 7: Factors impacting Collaboration Several factors affect collaboration. Some factors affect always positively and some do negatively while some others affect alternatively depending on other external conditions. There is a strong relationship between structural aspects of the SNGs and the effectiveness of collaborations. One main structural factor is heterogeneity, which is also described as the diversity / compatibility in resource profiles and social aspects of the collaborating SNGs. There are two kinds of heterogeneity: Type I heterogeneity is about differences in resources and capabilities; and Type II heterogeneity is about social dimensions such as vision, management approach, and governance system. Notably, heterogeneity (either Type I or Type II) augments the interdependency between the SNGs which thus improves the effectiveness of the collaboration. if(overlap(SNGi.(resources, capabilities), SNGj.(resources, ∨ overlap(SNGi.(social aspects), capabilities)) SNGj.(social aspects)) ≡ min) ⇒ max(heterogeneity, SNGi, SNGj) Another factor is complexity of collaboration. Usually business collaborations are task oriented and goal driven relationships. Collaborations usually begin on a small set of tasks wherein the collaborating SNGs evaluate each other’s capabilities, and then start collaborations on larger/complex set of tasks. So, as the complexity of the task(s) increases, the need for collaboration also increases. Complexity is normally characterized along two dimensions: the scope and the intensity of collaborations. Scope is described as the range of services / products that must be produced or managed through the collaboration, and intensity is the degree of direct involvement on the part of the SNGs. For instance, two SNGs collaborating to manufacture a new machine require a higher level of commitment in terms of resources and manpower than a single sale-campaign. An increase in the complexity of the collaboration needs greater coordination mechanisms, resource expenditures, and constant contact between the SNGs on a regular basis. Complexity of Collaboration = scope of collaboration + intensity of collaboration = task complexity + degree of involvement[for n collaborated tasks, then]

Dominance is the degree to which resources (both physical and informational) flow to and from a SNG, and thus reflects the power of the SNG compared to other SNGs in the total SNG network. It also acquires the popularity and visibility within the network. Dominance allows SNGs to obtain control over resources, and participate in collaborations. Dominance also helps in improving the extent to which a SNG is able to pursue its own goals. A dominant SNG attracts more preferential linkages and increases subservience of the collaborators. Likewise, a dominant collaborator enhances the stability of the collaboration. The extent to which a SNG is dominant in a network or collaborates with a dominant collaborator does have a positive impact on the collaboration. Dominant(SNGi) ⇒ (max(resF low, SNGi, Network)) ∧ max(resF low, Network, SNGi) ∧ max(infoF low, SNGi, Network) ∧ max(infoF low, Network, SNGi) Embeddedness is the degree to which a SNG is embedded within the network, and the degree of positive experiences a SNG has by taking part in a series of collaborations. Experience in collaborations increases SNGs’ chances of surviving and maintaining existing collaborations. It also provides new capabilities and administrative skills that facilitate managing new collaborations. Further, their position in the network increases their attractiveness to other SNGs and gives a good number of varied contacts. Conversely, the negative argument of embeddedness is that it makes SNGs open to shocks from entities outside of the SNG network or insulating them from resources that exist beyond the network. Hence, networks built on repeated ties can reduce the flexibility, lock in resources, and ultimately lead to inferior quality. Emb(SNGi) = collab_rating(SNGi) ∪ count(SNGi, collabs)

∪ risk(SNGi, external_shocks) Degree of resource overlap defines the extent to which other SNGs have similar resources. Overlap has considerable impact on the competitive intensity in the business market and confines individual SNGs in their economic action. Even though resource overlap actually increases competitive intensity and thus compels SNGs’ for improvement, it means a negative impact on the collaboration. Res_overlap(SNGi) = overlap(SNGi.resources, Network resources)

Reachability_overhead is the cost involved in reaching the prospective collaborator. Sometimes even if a collaborator is found, it has to be discarded because of excessive reachability overhead. Reachability overhead is a summation of cost of reaching each other and other risks of transportation such as vehicles availability, route availability, etc. Reachability_overhead(SNGi, SNGj) = reaching_cost (SNGi, SNGj) ∪ transportation_risk(SNGi, SNGj)

6. CONCLUSIONS AND PERSPECTIVES This paper formalizes a collaborations in SNGs. Tasks are also formalized and attributes are briefly defined. Terms such as goal, distance, subdistance, subtasks, collaborated distance, task-collaborator graph, and task classification are introduced and defined as per the scope of SNGs. Publish Subscribe Model is proposed and the collaboration process is described. Several constructs of communication language and various communication messages are also defined. The process of searching for a suitable collaborator was given. Also, collaboration rating is defined and the process of computation of collaboration rating is also described.

ACKNOWLEDGMENTS The author is grateful to Prof H Mohanty [DCIS, HCU] for pointing out the research direction explored in this paper as well as for helping in framing and presenting the theory presented here. This work is funded by Rajiv Gandhi National Fellowship (RGNF), India.

REFERENCES

[9]. X. Mi, and W. Shen,“Computer supported Collaborative Product Development: a Review”, Proceedings of CSCWD2005, Coventry, UK, 2005. [10]. Weigang Li, Wenbin Wang, Junyi Shen, On Key Issues in Information System for Collaborative Product Development Process, Proceedings of the 10th International Conference on Computer Supported Cooperative Work in Design, 2006 IEEE. [11]. Jianxin Li, Jinpeng Huai, Chunming Hu , PEACE-VO: A Secure Policy-Enabled Collaboration Framework for Virtual Organizations , 26th IEEE International Symposium on Reliable Distributed Systems 1060-9857/07, 2007 [12]. Danny Weyns, Tom Holvoet, Synchronous versus Asynchronous Collaboration in Situated Multi-agent Systems AAMAS’03, July 14 - 18, 2003, ACM. [13]. Sarosh N. Talukdar, Collaboration rules for autonomous software agents, Decision Support Systems, (1999), p269 - 278. [14]. Sreedhar Bhukya: A Social Network Model for Academic Collaboration. NDT 2011: 203-211 [15]. Assessing the Value of Collaborations in Tourism Networks: A Case Study of Elkhart County, Indiana. Florian Zach, Pradeep Racherla. JOURNAL OF TRAVEL & TOURISM MARKETING VOLUME 28, ISSUE 1, 2011 .

[1]. Sousa, C., Martins, P., Fonseca, B., Paredes, H., Meehan, A., Devine, T.: Social networking system for academic collaboration. In: Luo, Y. (ed.) CDVE 2008. LNCS, vol. 5220, pp. 295–298. Springer, Heidelberg (2008)

[16]. Collaborative Workflow Solution for Distributed Product Development. Yuan-Ping Luh, ChihChin Pan. . Int. J. Computer Integrated Manufacturing 24(5): 471-483 (2011).

[2]. Hill, V.A.: Collaboration in an Academic Setting: Does the Network Structure Matter? NSF 07-576 (CASOS)

[17]. “WfMC, Workflow management Coalition”, Workflow reference model, 1995, http://www.wfmc.org.

[3]. Watts, D.J., Strogatz, S.H.: Collective dynamics of ‘small -world’ networks. Nature 393, 440 (1998)

[18]. Liu, J., Zhang, S. & Hu, J., “A case study of an inter-enterprise workflow-supported supply chain management system”, Information and Management, 2005, 42 (3), 441-454.

[4]. Newman, M.: The structure of scientific collaboration networks. PNAS 98, 404–409 (2001) [5]. Newman, M.: Coauthorship networks and patterns of scientific collaboration. PNAS 101, 5200–5205 (2004) [6]. Jihye Song, Minjoon Kim, Haksung Kim, Kwanghoon Kim. A Framework: Workflowbased Social Network Discovery and Analysis. 13th IEEE International Conference on Computational Science and Engineering, 2010. [7]. Holme, P., Edling, C.R., Liljeros, F.: Structure and Time-Evolution of an Internet Dating Community. Soc. Networks 26, 155–174 (2004) [8]. Obinna Anya, Hissam Tawfik and Atulya Nagar, A Conceptual Design of an Adaptive and Collaborative e-Work Environment, Proceedings of the First Asia International Conference on Modelling & Simulation, (AMS’07).