Ontology development and optimisation for data integration and decision-making in product design and obsolescence management

Ontology development and optimisation for data integration and decision-making in product design and obsolescence management

‘Ontology Modelling in Physical Asset Management‘. From the introduction to Chapter 5: ‘On Ontology Development and Optimisation for Data Integration and Decision Making in Product Design and Obsolescence Management’.

Product design affects most of the immediate costs incurred for the product (e.g. materials, time, labour) and the costs incurred over the life of the product (e.g. manufacturing, maintenance, distribution, service, disposal). It is therefore important to obtain cost estimates as early and as quickly as possible at the design stage (Jo et al. 1993).

Furthermore, 80-90% of time-to-market is related to product planning and development due to product complexity (Charney 1991) and 40% of all quality problems are attributed to design defects (Saaksvuori and Immonen 2005).

If companies aim to significantly reduce time-to-market and increase product quality, they must address efficiency and quality in product design.

On the other hand, low-volume products and systems, such as military and avionics, often use commercially available high-tech components, and in the past decade, technology has advanced so fast that such components have a short procurement life and quickly become obsolete. However, low-volume custom products and systems, such as submarines and aircraft, can remain in use for decades (Josias et al. 2004; Sandborn 2008). Many components become obsolete before the system is put into service. Component obsolescence poses significant problems and high costs for the manufacture and maintenance of complex systems with a long field life, and such obsolescence is referred to as Diminishing Manufacturing Sources and Material Shortages (DMSMS), where the original manufacturer, meaning the loss of the ability to source components and technology from the original manufacturer).

Computer-based approaches have been developed and used by designers to reduce design time, reduce product costs and improve product quality. Several CAX (Computer Aided X) and DFX (Design for X) tools have been developed for this purpose over the past two decades.

The common goal of all Design for Manufacturing (DFM) systems will be to minimise the total cost of the product life cycle by making more systematic and efficient decisions when considering design with manufacturing in mind (O’Flynn and Ahmad 1993 ).

With the impact and spread of DMSMS obsolescence, a number of tools have been developed and used to manage this issue. Examples include Raytheon’s Component Obsolescence and Reuse Tool (CORT), IHS’s COMETTM, QinetiQ’s Q-StarTM, IGG’s Obsolescence Monitoring Service and ARINC’s Obsolescence Management Programme.

The most prominent obsolescence management initiative is the US Department of Defence’s (DoD) Diminishing Manufacturing Sources and Material Shortages (DMSMS) Knowledge Sharing Portal (KSP) (McDermott 2002), where KSPs are used by government agencies, industry and supplier representatives to provide a forum for identifying and resolving obsolescence-related issues.

These tools are database-based and feature database management of information on obsolete parts, which can report on the current obsolescence status of a part, predict the risk of obsolescence and identify possible replacement parts.

Existing systems have provisions for information and data to be produced during product design and obsolescence management, but the complete set of information needed for decision-making may not be available, consistent or expressed in a common format (Zheng et al. 2013), A study by Coopers and Lybrand reported that 59% of time spent on design could be saved if information systems enabled information to be shared, retrieved and re-used efficiently.

Furthermore, in many organisations, product development involves a wide range of stakeholders, including designers, scattered across time and space and across the globe, using multiple computer-aided tools. Information sources are stored in isolated and heterogeneous data resources, making it a challenging task to integrate information from data resources for optimal design.

In obsolescence management, dispersed and heterogeneous sources of information hinder the introduction of supply chain knowledge, such as demand and inventory integration, in the application of current tools, and therefore existing research on obsolescence in DMSMS focuses on managing obsolescence at the component level in a reactive manner, i.e. after problems have occurred and The focus has been on minimising the cost of resolving them.

A common root cause of today’s information systems is the lack of a know-how edge model to support data integration and decision-making in product design and obsolescence management, and there are several technical and environmental factors that must be considered simultaneously in product design, which are interrelated and and designers need to support product design decisions in the face of a wealth of data integrated from multiple data resources and the complex relationships between them.

For product obsolescence, DMSMS solutions need to move from reactive management at the component level to proactive and strategic management at the product and enterprise level in order to resolve DMSMS obsolescence at a lower total cost.

Knowledge representation, which enables information sharing, re-use and cooperation on design and obsolescence issues between different organisations, is the basis for decision-making in product design and proactive and strategic management of DMSMS obsolescence, and to establish a comprehensive knowledge representation scheme, the backbone of the required information model ontology is needed.

In recent years, the concept of ontology has been used in the fields of knowledge management and Computer Supported Cooperative Work (CSCW). An ontology is a formal specification of domain knowledge and is used to define a set of data and its structure for experts to share information in a domain of interest, and ontologies are good at representing and exploiting relationships between data and can make knowledge inference efficient (Noy and McGuiness 2001).

This ontology is useful for integrating heterogeneous data. This is because the problem of heterogeneity in data integration is very common in distributed information systems (Kim and Seo 1991; Kashyap and Sheth 1997).

Heterogeneity problems have been classified into three categories: syntax (data format heterogeneity), structure (homonyms, synonyms and different attributes in database tables) and semantics (the intended meaning of terms in special contexts and applications) (Zheng 2011 ).

In recent years, systems have emerged that utilise the benefits of ontologies to enable information reasoning and data retrieval, and ontology-based data integration can direct users to the appropriate data source or department, retrieve heterogeneous data from related data sources (Chang and Terpenny 2008a, b; 2012), transforming data into the required format (Chang et al. 2009) and resolving data conflicts (data with different formats and representations) from different sources (Zheng et al.).

Ontologies can also help decision-making in product design and obsolescence management, taking into account complex relationships and constraints.

For example, the KSL Wine Agent (http://www.ksl.stanford.edu/people/dlm/webont/wineAgent/) developed by McGuinness et al. is a testbed application for Semantic Web technologies, and has been used in the JTP ( Java Theorem Prover) (Frank et al. 2004), which combines a logical reasoning system and an object-oriented modular reasoning system with the OWL (Web Ontology Language) ontology.

After receiving the description of the meal, the wine agent searches for a selection of matching wines available on the web. This is achieved by referencing the ontology, executing the query and outputting the results.

Thus, as in the wine industry, where restaurants, customers and retailers exist, Semantic Web technology can also be used for product design and obsolescence management, which relates products, components, manufacturing processes, machinery, materials, inventory, suppliers, etc.

The wine agent methodology can be applied to build ontology-based information systems for product design and obsolescence management.

For example, engineers can mark each product with a machine-readable standard definition, manufacturers can mark manufacturing activities, providers can mark the resources required by those activities, and to reduce the obsolescence impact of the DMSMS as much as possible, strategic managers can use the parts Based on the forecast of the procurement lifecycle, sources of components that are not likely to become obsolete in the near future can be marked for design and production.

Instead of the traditional approach of trying to build up a huge database of products, resources and activities, the use of concepts connected by ontologies, distributing definitions across participants and utilising semantic relationships within the ontology, allows the appropriate information to be found after querying the ontology can be found.

This means, for example, that cost data can be retrieved through a series of queries from item costs to activity costs and finally to resource costs, including relevant information such as item, activity, material and resource details. The ontology can also be queried to obtain suggestions on design and obsolescence management, such as wine selection suggestions in wine agents.

This chapter details the development, optimisation, maintenance and utilisation of ontologies for products. The advantages of ontologies in knowledge representation and reasoning are exploited using ontology-based data integration systems to search and integrate heterogeneous data resources, input data into analytical models and support decision-making with the integrated data.

They have also built ontology-based decision support tools using semantic rule languages and reasoning tools to adjust product design parameters, calculate obsolescence mitigation costs, and provide decision support in design strategy and obsolescence management.

Examples of implementing systems for data integration and decision-making in product design and obsolescence management using ontologies

The following section describes an example implementation of an ontology-based data integration and decision-making system in product design and obsolescence management.

Role of ontologies: an ontology is a formal specification that defines concepts and their relationships in a particular domain, and in product design and obsolescence management ontologies play the following roles

1. data integration: providing a framework for understanding and relating data from different sources in a unified way

2. knowledge sharing: improving communication between teams by defining common terms and meanings

3. decision support: supports decision-making using inference engines by modelling knowledge in a format that can be interpreted by computers.

Example implementations:.

1. build a product design ontology:

– Purpose:  to model product characteristics, design processes and user requirements.
– Steps:
1. identify the main concepts in the domain (e.g. parts, functions, requirements).
2. define the relationships between these concepts (e.g. parts provide functions, requirements influence the design process).
3. describe the ontology using OWL (Web Ontology Language) or RDF (Resource Description Framework). See also “Ontology technology” for more information.

2. data integration and analysis:

– Purpose: to integrate and analyse data from different sources (e.g. market research, user evaluations, product specifications).
– Tools:
– Ontology-based data querying using SPARQL, also described in “About RDF stores and SPARQL“.
– Collect and integrate data using data integration platforms (e.g. Apache Jena, Protégé). 3.

3. support for obsolescence management:

– Purpose: to optimise product longevity and reduce the risk of obsolescence.
– Processes:
1. collect life cycle data on product components.
2. modelling the relationships and dependencies of the parts using an ontology.
3. use an inference engine (e.g. Apache Jena) to assess obsolescence risk and suggest alternative parts. For more information on inference techniques, see also “Inference techniques“.

4. development of decision support systems:

– Purpose: to assist designers and managers to make decisions quickly.
– Functions:
– Develop ontology-based visualisation tools to visually display data relationships.
– Scenario analysis to predict the impact of different design options.

The benefits of these systems include.

– Efficient data management: ontologies organise complex data and make them easily accessible.
– Improved decision-making: providing insight based on accurate data analysis and reasoning.
– Improved collaboration: effective communication between different teams through a common knowledge base.

Reference Information and Reference Books

Detailed information on the use of knowledge graphs can be found in “Knowledge Information Processing Technology” “Ontology Technology” “Semantic Web Technology” and “Inference Technology. Please refer to them as well.

Also, reports from academic conferences, such as those described in the “Collection of AI Conference Papers” are also helpful.

Reference book is “Building Knowledge Graphs“

“Knowledge Graphs and Big Data Processing“

“The Knowledge Graph Cookbook“

“Domain-Specific Knowledge Graph Construction“