Research on Discovery and Evaluation of Network Manufacturing Resources

Libin Zheng, Jinan Gu2015 · DOI

Abstract

For the disadvantages of discovery mechanism of current networked manufacturing resources based on Semantic Web Service, a joint discovery mechanism based on UDDI (Universal Description, Discovery and Integration) and WSIL (Web Services Inspection Language) was put forward. Basic process features are matched in UDDI and corresponding eigen values are matched in local WSIL. An evaluation model combined fuzzy theory with analytic hierarchy process and the law of comparative judgment was presented. Discovery architecture of networked manufacturing resources was built. Basic principles and characteristics of the integrated architecture were analyzed. Key technologies for realizing the integrated architecture were studied. The joint discovery mechanism prototype platform for manufacturing resource was developed by borrowing ideas from this model. The system improves the efficiency of finding manufacturing resources and enhances scientificalness and reliability of evaluation.

Introduction

In an era of economic globalization, core enterprises frequently find their internal production capacity insufficient for specific manufacturing tasks. To bridge this gap, they must search for external manufacturing resources. However, existing discovery mechanisms face significant challenges. Centralized systems like UDDI are manageable but suffer from heavy server loads as resource volume grows. Conversely, distributed P2P systems avoid bottlenecks but are notoriously difficult to search and manage. Furthermore, traditional keyword matching is often too rigid; if a service provider uses a synonym or a slightly different term than the seeker, the resource remains hidden. Once a resource is found, evaluating its suitability is equally complex, as human judgment often becomes less accurate when comparing many factors simultaneously.

Proposed Joint Discovery Mechanism

To address these hurdles, this research introduces a multi-layered joint discovery mechanism. The system uses UDDI to store basic process features (such as "shaft feature" or "hole feature") and WSIL to manage detailed local constraints (such as specific machining accuracy or size limits) within the enterprise's own network.

Integrated architecture for resource discovery, illustrating the workflow from modeling to local constraint matching.

As illustrated in the architecture above, the process begins with resource modeling and publishing. High-level feature matching occurs in the UDDI registry. If a match is found, the system links to the local enterprise's WSIL for feature constraint matching. This hybrid approach balances centralized management ease with decentralized performance.

Semantic Expansion and Hybrid Evaluation

To improve retrieval flexibility, the system employs Domain Ontology to calculate "semantic similarity" between concepts. If the similarity score between a requested feature and a listed resource exceeds a set threshold, the system considers them a match, even if the terms are not identical.

For ranking discovered resources, the authors present a hybrid evaluation model combining Fuzzy Theory, Analytic Hierarchy Process (AHP), and the Law of Comparative Judgment. Recognizing that experts are often uncertain, the model utilizes interval estimation for weights rather than single numbers. It also incorporates the expert's risk profile (averse, neutral, or appetite) into the mapping function to refine final weights, ensuring a more reliable and scientific selection process.

Implementation and Results

The authors developed a prototype platform using XML, Visual Studio.NET, and SQL Server 2000 to verify the architecture. The evaluation utilized simulated equipment resource data, including machine tool types such as the CAK1626 and CAK3665.

Prototype results showing discovered manufacturing resources and their technical parameters.

As shown in the results table, the system successfully filtered resources based on location (e.g., Shanghai, Beijing), machining accuracy (IT6, IT7), and availability status. The semantic matching engine effectively identified synonymous manufacturing capabilities that traditional keyword searches would miss, significantly improving retrieval efficiency.

Conclusion

The joint UDDI/WSIL discovery mechanism provides a flexible and efficient framework for identifying networked manufacturing resources. By embracing uncertainty in evaluation through fuzzy-AHP and accounting for expert risk attitudes, the model reflects the practical reality of industrial decision-making. Future work will focus on modernizing the technology stack and conducting large-scale benchmarking to ensure performance as the network of manufacturing nodes continues to expand.

Summary updated

Abstract

For the disadvantages of discovery mechanism of current networked manufacturing resources based on Semantic Web Service, a joint discovery mechanism based on UDDI (Universal Description, Discovery and Integration) and WSIL (Web Services Inspection Language) was put forward. Basic process features are matched in UDDI and corresponding eigen values are matched in local WSIL. An evaluation model combined fuzzy theory with analytic hierarchy process and the law of comparative judgment was presented. Discovery architecture of networked manufacturing resources was built. Basic principles and characteristics of the integrated architecture were analyzed. Key technologies for realizing the integrated architecture were studied. The joint discovery mechanism prototype platform for manufacturing resource was developed by borrowing ideas from this model. The system improves the efficiency of finding manufacturing resources and enhances scientificalness and reliability of evaluation.
BibTeX
@misc{zheng2015research,
  author = {Libin Zheng and Jinan Gu},
  title = {{Research on Discovery and Evaluation of Network Manufacturing Resources}},
  year = {2015},
  doi = {10.2991/icadme-15.2015.91},
}
Figure