Experience capitalization to support decision making in inventive problem solving
Introduction
Innovation in the workplace has become an increasingly important determinant of organizational performance, success, and longer-term survival [1], [2]. The generation of creative ideas is the first stage activity that leads to innovation. Creative ideas are specific inventive technical solutions to the specific inventive technical problems (hereafter, we refer the former as specific solutions and the latter as specific problems).
The theory of inventive problem solving (TRIZ) [3] is recognized as a suitable methodology to facilitate the resolution of specific problems through the use of different tools, models and knowledge sources [4], [5]. According to the TRIZ methodology, solving a specific problem goes through three phases [6] as illustrated in Fig. 1.
- 1.
The “formulation” phase, where the experts formulate their specific problem into an abstract problem in terms of a physical contradiction, a technical contradiction or a substance-field model.
- 2.
The “abstract solution finding” phase, where, depending on the type of the abstract problem obtained in the former phase, different knowledge sources are used to get one or more abstract solutions.
- 3.
The “interpretation” phase, where the experts interpret the abstract solutions using other specific knowledge sources and their own creativity in order to get one or more specific solutions.
TRIZ is based on the use of tools (e.g. the Contradiction Matrix,1 the Substance-Field analysis2 and the use of the pointers to the scientific effects3, etc.) to build a certain type of abstract problem. It is also based on the use of different knowledge sources (e.g. 40 inventive principles,4 the 76 inventive standards,5 the scientific effects, etc.) to find abstract solutions. However, the TRIZ knowledge sources are built independently from each other at different levels of abstraction. Depending on the abstraction level, they produce results also at different levels: at the highest abstraction level, the results are ideas of solutions (abstract solutions), at the lowest abstraction level, they produce specific solutions. The use of TRIZ for specific problem solving requires a high level of expertise. Indeed:
- •
TRIZ's body of knowledge are sometimes ambiguous and cannot be interpreted adequately. The methodology does not currently provide clear guidelines for the use of its knowledge sources to solve a specific problem.
- •
TRIZ provides the wealth of knowledge for solving a large variety of inventive problems, but the access to the specific needed knowledge might be troublesome.
In order to facilitate the use of TRIZ, several works have been carried out using ontologies. The authors in [7] explored the indexing of knowledge involved in the patents. A TRIZ Technical System ontology is developed in this work and an ontology-based classification methodology is implemented to support the retrieval of related information extracted from patents. Moreover, the work in [8] explored the formulation of two of TRIZ knowledge sources based on ontology. Recently, a rule-based approach [9] has found its way to facilitate the search for abstract solutions using an ontology and rules, as it is illustrated in Fig. 2(a). Firstly, it formalized two TRIZ knowledge sources (40 Inventive Principles, 76 Inventive Standards) with different abstraction levels based on ontologies. Then, the knowledge base was populated with the TRIZ knowledge sources and its associated rules. In this way, each rule specifies the representation of different pieces of knowledge involved in the TRIZ knowledge sources and enables reasoning to search for abstract solutions.
However, the use of the rule-based approach presents some disadvantages, mainly associated with the fact that the introduction of additional knowledge to solve a specific problem might make the whole model logically incoherent. This is exactly the case when an adaptation of an abstract solution can be applied to solve similar specific problems, because new specific knowledge needs to be included in the original knowledge base to calculate this adaptation. Additionally, the use of a rule-based approach for solving problem implies that every new specific problem needs a new solving process, making the whole time consuming.
To cope with the drawbacks of the rule-based approach, we are interested here in finding ways to reuse existing solutions to solve new specific problems. Therefore, there is a need to capitalize experience through collecting and making use of previous cases in order to enhance problem solving [10]. In this paper, we propose a new approach for inventive problem solving through experience capitalization. The contribution of our approach is twofold: (1) it helps the new TRIZ users to solve inventive problems more easily by avoiding going through the TRIZ problem solving process; (2) it helps the users to obtain the solution directly from capitalized experiences, thus reducing the time needed for inventive problem solving. There are several ways to facilitate the capitalization of experience [11], [12]. Among them, case-based reasoning (CBR) is relevant to our needs. In particular, we built the case base to facilitate the storage of the additional needed knowledge without risking the logical coherence of the whole model. The case base is implemented as an ontology in order to store different types of cases represented by different features. Furthermore, specific problems can be solved by finding and reusing old solutions. In this way, solving a specific problem is now possible by adapting old specific solutions, as depicted in Fig. 2(b). The user finds the old specific problems that are similar to the new one based on a similarity calculation. The new specific problem can be solved by reusing or revising the old specific solutions of the obtained similar old problems. If the CBR approach fails (because, for example, there are no similar old problems or adaptation of old solutions is not possible), the classic rule-based approach is used.
The remainder of the paper is organized as follows. Section 2 presents the related concepts and the literature review about different approaches to facilitate the resolution of specific problems. Section 3 describes the general architecture of the proposed approach. In Section 4, the proposed approach is described in detail, including the two modules to facilitate specific problem solving. In Section 5, the resolution of a specific problem is given to illustrate the proposed approach. Section 6 introduces the implementation and evaluation of the proposed approach. Finally, Section 7 concludes with a summary and highlights some directions for future work.
Section snippets
The theory of inventive problem solving
TRIZ is the Russian acronym for the Theory of Inventive Problem Solving [3]. The basic idea of TRIZ is that there are general principles that can be used as abstract solutions. The abstract solutions can be applied to similar problematic situations by adaptations. Therefore, TRIZ is used to solve inventive problems by going through three phases: formalizing the abstract problem, finding the abstract solution and then interpreting the abstract solution to the specific solution according to the
General architecture of the proposed approach
In this section, we present the general architecture of the proposed approach (Fig. 4). On the one hand, it adds an alternative route to the classical TRIZ problem solving by solving the specific problem directly. On the other hand, it uses one or several modules independently or collectively to facilitate the resolution of the specific problem. Generally speaking, the proposed approach contains four modules:
- •
The Experience-based specific solution finding module (ESS) is proposed to provide an
Description of the proposed approach
In this section, we give a detailed description of the ESS and EAP modules.
Case study
To better illustrate our approach, we present in this section a case study related to cloth hangers. We describe in detail how this specific problem is resolved by going through the modules presented in Section 4. To test the approach, we consider a case base of 47 cases that have been solved by engineering students and experts.
Implementation
The prototype10 implementing the proposed approach, named CBRID (case-base reasoning for inventive design), is developed as a Java EE web application, using the Spring Boot framework version 1.5.9 11. It is deployed on a Tomcat12 server version 8.
Evaluation
The ISO 9241-11 standard13 defines usability as “extent to which a product can be used by specified users to
Discussion and conclusion
This paper explores an approach for experience capitalization for inventive problem solving. Within this paper, we only deal with two modules that applied CBR to facilitate the capitalization of experience, the ESS module and the EAP module. A specific solution can be obtained primarily by reusing or revising one of the cases retrieved in the ESS module through the use of different similarity calculation algorithms. If the similar specific problem is not found or the revision based on the old
References (51)
- et al.
The development of a modified TRIZ technical system ontology
Comput. Ind.
(2012) - et al.
An ontology-based approach for inventive problem solving
Eng. Appl. Artif. Intell.
(2014) - et al.
On the complementarity of TRIZ and axiomatic design: from decoupling objective to contradiction identification
Proc. Eng.
(2011) - et al.
Grouping of TRIZ inventive principles to facilitate automatic patent classification
Expert Syst. Appl.
(2008) - et al.
A method for facilitating inventive design based on semantic similarity and case-based reasoning
Proc. Eng.
(2015) - et al.
Concept learning and heuristic classification in weak-theory domains
Artif. Intell.
(1990) - et al.
A CBR system for injection mould design based on ontology: a case study
Comput.-Aided Des.
(2012) A translation approach to portable ontology specifications
Knowl. Acquisit.
(1993)- et al.
IngeniousTRIZ: an automatic ontology-based system for solving inventive problems
Knowl.-Based Syst.
(2015) - et al.
An ontology-based approach for inventive problem solving
Eng. Appl. Artif. Intell.
(2014)
How to make a decision: the analytic hierarchy process
Eur. J. Oper. Res.
Innovation and creativity in organizations: a state-of-the-science review, prospective commentary, and guiding framework
J. Manage.
Innovation Based Competition & Design Systems Dynamics: Lessons From French Innovative Firms and Organizational Issues for the Next Decade
The Innovation Algorithm: TRIZ, Systematic Innovation and Technical Creativity
Industry best practices and the role of TRIZ in developing new products
International Research Conference ”TRIZFest
Creativity as an Exact Science: The Theory of the Solution of Inventive Problems
(Meta) Knowledge Modeling for Inventive Design, (Ph.D. thesis)
A hybrid system combining description logics and rules for inventive design
KREM: a generic knowledge-based framework for problem solving in engineering
KEOD
Case-based reasoning: foundational issues, methodological variations, and system approaches
AI Commun.
An OWL ontology of set of experience knowledge structure
J. UCS
Cited by (13)
SMFM-based analogy retrieval tool for the conceptual design of innovative products
2023, Computers in IndustryA new computational method for acquiring effect knowledge to support product innovation
2021, Knowledge-Based SystemsCitation Excerpt :The technical characteristic implicit in category information of main groups and subgroups, generally relevant to functionality and principle or structure, can be extracted as design knowledge [20]. Classification numbers or category information of section-class-subclass are competent for the technical area definitions of products/design problems [21,22]. To acquire the desired knowledge from IPC, our method first pre-processes IPC text reasonably, then extracts technical area and functional features of effects respectively by regular expression matching and a technology combining syntactic analysis and word vector (i.e., the global vector [23], abbreviated as GloVe), and finally adopts a taxonomy consisting of an existing functional taxonomy (i.e., Functional Basis [24]) and the section-class-subclass taxonomy of IPC, and concept classification methods, such as WordNet-based [25] word similarity measures and GloVe [23], to classify the effects extracted as effect-function-flow-technical area knowledge.
Case-based Reasoning for crisis response: Case representation and case retrieval
2020, Procedia Computer ScienceKnowledge-Driven Industrial Intelligent System: Concept, Reference Model, and Application Direction
2023, IEEE Transactions on Computational Social SystemsEmergency Management Case-Based Reasoning Systems: A Survey of Recent Developments
2023, Journal of Experimental and Theoretical Artificial Intelligence