Sun et al. (2016) proposed identifying a collaborative team — a leader (host enterprise) and partner enterprises — for an engineering project on a weighted social network of manufacturing service providers mined from historical workflow logs, using a betweenness-based leader-selection algorithm (MLS) and a connector-node-restricted member-discovery algorithm (DCAM). Validation was limited to static simulations of 40–280 providers with fixed task requirements, and the method carries no quantitative notion of trust between providers. The open problem, stated in the source's conclusions and left unresolved by subsequent literature, is to extend this framework into a trust-aware and adaptive one: integrate quantitative trust metrics into leader and member selection, adapt to dynamic evolution of the provider network and fluctuating service availability, and demonstrate effectiveness for large-scale, heterogeneous provider populations beyond the simulation-only setting, toward real-world manufacturing collaboration.

Background

In service-oriented manufacturing, an engineering project is often carried out not by a single enterprise but by a temporary alliance: one enterprise acts as the host (the leader) and other service providers contribute the required capabilities. Sun et al. (2016) model such providers as a weighted social network: from historical workflow (business-process) event logs they compute, for each pair of services, a dependency support (how often the two services co-occur in process instances) and a confidence (the conditional probability that one is directly followed by the other); an edge between two providers is created if some service pair of theirs shows a dependency, and its weight is derived from the dependency confidence over all cross-pairs of their services, with a lower weight read as more effective past collaboration. Given a project's set of required tasks, the team-formation problem is to choose a leader and providers covering all tasks so as to minimize the leader distance, i.e. the sum of shortest-path distances in this weighted graph from the leader to each chosen member. This connects to the team-formation-with-a-leader line in social networks (Lappas, Liu and Terzi 2009; Kargar and An 2011), where exhaustively enumerating candidate teams is computationally prohibitive; Juang, Huang and Huang (2013) gave pruning algorithms (BCPruning and SSPruning) for the leader case. Sun et al. accelerate leader selection with an algorithm called MLS: an enhanced graph is built by attaching an artificial node for each required task, connected with a large constant weight to every provider able to perform that task, and leaders are ranked by a LeaderRank — a betweenness-centrality-style score (betweenness centrality, computable by Brandes' 2001 algorithm in O(nm+n2log⁡n)O(nm+n^{2}\log n) time on a graph with nn vertices and mm edges) in which only the artificial task nodes serve as source vertices, reducing the cost to O(Tnum⋅m+Tnum⋅nlog⁡n)O(T_{num}\cdot m+T_{num}\cdot n\log n) with TnumT_{num} the number of required tasks. Member discovery (the DCAM algorithm) then searches only connector nodes — providers in a task's functional group (the providers capable of that task) that have a neighbor in a different functional group — which provably suffices for shortest paths arriving from a leader outside the group. Validation was by simulation only: networks of 40–280 providers and 2–12 required tasks, showing lower computation time than brute-force enumeration, BCPruning and SSPruning while keeping leader distance close to the exact solution. Two families of limitations motivate the open question. First, trust: the edge weights encode past co-occurrence patterns, not trustworthiness, even though quantitative trust for provider selection in social networks was already established (the optimal social trust path of Liu, Wang, Orgun and Lim 2013; dynamic group trust for service selection in Wei, Zheng and Chen 2014), and the same authors had earlier built trust into workflow scheduling in cloud computing (Tan et al. 2014) — but not into partner identification. Second, dynamics and scale: the method assumes a static graph, fixed service availability, and fixed task requirements, and was tested only at up to 280 nodes, whereas real manufacturing ecosystems evolve, providers and services fluctuate in availability, and populations are large and heterogeneous.

Problem Statement

As stated in the conclusions of Sun et al. (2016), the unresolved problem is to extend their collaborative partner identification framework so that it becomes usable in live, real-world manufacturing collaboration. Concretely, the open question asks for a partner-identification method over social service provider networks that: (1) integrates quantitative trust metrics into the team-formation process, so that leader selection and member discovery account for the measured trustworthiness of providers rather than only past-collaboration cost (leader distance); (2) is adaptive, responding to the dynamic evolution of the social network (collaboration relationships forming and dissolving) and to fluctuating service availability, rather than assuming a fixed static graph with fixed task requirements; and (3) is validated on large-scale, heterogeneous provider populations beyond the original simulations of at most 280 providers, with real-world manufacturing applications as an aspirational extension rather than a prerequisite for resolution. These are coupled aspects of one objective — making workflow-mined, social-network-based partner identification deployable and dependable in evolving manufacturing ecosystems — not independent deliverables.

References

  1. Yong Sun, Wenan Tan, Lingxia Li, Weiming Shen, Zhuming Bi, Xiaoming Hu (2016). A new method to identify collaborative partners in social service provider networks. Information Systems Frontiers 18(3):565–578. DOI: 10.1007/s10796-015-9547-4. https://doi.org/10.1007/s10796-015-9547-4
  2. Ming-Chin Juang, Chen-Che Huang, Jiun-Long Huang (2013). Efficient algorithms for team formation with a leader in social networks. The Journal of Supercomputing 66(2):721–737. DOI: 10.1007/s11227-013-0907-x. https://doi.org/10.1007/s11227-013-0907-x
  3. Theodoros Lappas, Kun Liu, Evimaria Terzi (2009). Finding a team of experts in social networks. In Proceedings of the 15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD '09), pp. 467–476. DOI: 10.1145/1557019.1557074. https://doi.org/10.1145/1557019.1557074
  4. Mehdi Kargar, Aijun An (2011). Discovering top-k teams of experts with/without a leader in social networks. In Proceedings of the 20th ACM International Conference on Information and Knowledge Management (CIKM '11), pp. 985–994. DOI: 10.1145/2063576.2063718. https://doi.org/10.1145/2063576.2063718
  5. Ulrik Brandes (2001). A faster algorithm for betweenness centrality. Journal of Mathematical Sociology 25(2):163–177. DOI: 10.1080/0022250X.2001.9990249. https://doi.org/10.1080/0022250X.2001.9990249
  6. Guanfeng Liu, Yan Wang, Mehmet A. Orgun, Ee-Peng Lim (2013). Finding the optimal social trust path for the selection of trustworthy service providers in complex social networks. IEEE Transactions on Services Computing 6(2):152–167. DOI: 10.1109/TSC.2011.58. https://doi.org/10.1109/TSC.2011.58
  7. WenAn Tan, Yong Sun, Ling Xia Li, GuangZhen Lu, Tong Wang (2014). A trust service-oriented scheduling model for workflow applications in cloud computing. IEEE Systems Journal 8(3):868–878. DOI: 10.1109/JSYST.2013.2260072. https://doi.org/10.1109/JSYST.2013.2260072
  8. Yong Sun, Wenan Tan, Quanquan Zhang (2016). An efficient algorithm for crowdsourcing workflow tasks to social networks. In 2016 IEEE 20th International Conference on Computer Supported Cooperative Work in Design (CSCWD), pp. 532–538. DOI: 10.1109/CSCWD.2016.7566046. https://doi.org/10.1109/CSCWD.2016.7566046
  9. Yong Sun, Wenan Tan, Li Huang (2017). A team discovery model for crowdsourcing tasks to social networks. International Journal of Web Engineering and Technology 12(1). DOI: 10.1504/IJWET.2017.084023. https://doi.org/10.1504/IJWET.2017.084023
  10. Yong Sun, Wenan Tan (2019). A trust-aware task allocation method using deep Q-learning for uncertain mobile crowdsourcing. Human-centric Computing and Information Sciences 9(1):25. DOI: 10.1186/s13673-019-0187-4. https://doi.org/10.1186/s13673-019-0187-4
  11. Lu Zhao, Wenan Tan, Na Xie, Li Huang (2020). An optimal service selection approach for service-oriented business collaboration using crowd-based cooperative computing. Applied Soft Computing 92:106270. DOI: 10.1016/j.asoc.2020.106270. https://doi.org/10.1016/j.asoc.2020.106270
  12. Na Xie, Wenan Tan, Xianrong Zheng, Lu Zhao, Li Huang, Yong Sun (2021). An efficient two-phase approach for reliable collaboration-aware service composition in cloud manufacturing. Journal of Industrial Information Integration 23:100211. DOI: 10.1016/j.jii.2021.100211. https://doi.org/10.1016/j.jii.2021.100211
  13. Yash Mahajan, Jin-Hee Cho, Ing-Ray Chen (2024). Privacy-preserving and diversity-aware trust-based team formation in online social networks. ACM Transactions on Intelligent Systems and Technology 15(5):1–32. DOI: 10.1145/3670411. https://doi.org/10.1145/3670411
  14. Matthew E. Gaston, Marie desJardins (2005). Agent-organized networks for dynamic team formation. In Proceedings of the Fourth International Joint Conference on Autonomous Agents and Multiagent Systems (AAMAS '05), pp. 230–237. DOI: 10.1145/1082473.1082508. https://doi.org/10.1145/1082473.1082508
  15. Shouxian Wei, Xiaolin Zheng, Deren Chen (2014). Service selection based on dynamic group trust in social network. In 2014 International Conference on Service Sciences (ICSS), pp. 157–162. DOI: 10.1109/ICSS.2014.38. https://doi.org/10.1109/ICSS.2014.38