This paper addresses the fulfillment of requirements related to case-basedreasoning (CBR) processes for system design. Considering that CBR processes are well suited for problem solving, the proposed method concerns the definition of an integrated CBR process in line with system engineering principles. After the definition of the requirements that the approach has to fulfill, an ontology is defined to capitalize knowledge about the design within concepts. Based on the ontology, mod- els are provided for requirements and solutions representation. Next, a recursive CBR process, suitable for system design, is provided. Uncertainty and designer preferences as well as ontological guidelines are considered during the requirements definition, the compatible cases retrieval, and the solution definition steps. This approach is designed to give flexibility within the CBR process as well as to provide guidelines to the designer. Such questions as the following are conjointly treated: how to guide the designer to be sure that the requirements are correctly defined and suitable for the retrieval step, how to retrieve cases when there are no available similarity measures, and how to enlarge the research scope during the retrieval step to obtain a sufficient panel of solutions. Finally, an example of system engineering in the aeronautic domain illustrates the proposed method. A testbed has been developed and carried out to evaluate the performance of the retrieval algorithm and a software prototype has been developed in order to test the approach. The outcome of this work is a recursive CBR process suitable to engineering design and compatible with standards. Requirements are modeled by means of flexible constraints, where the designer preferences are used to express the flexibility. Similar solutions can be retrieved even if similarity measures between features are not available. Simultaneously, ontological guidelines are used to guide the process and to aid the designer to express her/his preferences.
Cambridge, Massachusetts 02139
{beenkim, rudin, julie a shah}@csail.mit.edu
Abstract
We present the Bayesian Case Model (BCM), a general framework for Bayesian case-basedreasoning (CBR) and prototype classification and clustering. BCM brings the intuitive power of CBR to a Bayesian generative framework. The BCM learns prototypes, the “quintessential” observations that best represent clusters in a dataset, by performing joint inference on cluster labels, prototypes and impor- tant features. Simultaneously, BCM pursues sparsity by learning subspaces, the sets of features that play important roles in the characterization of the prototypes. The prototype and subspace representation provides quantitative benefits in inter- pretability while preserving classification accuracy. Human subject experiments verify statistically significant improvements to participants’ understanding when using explanations produced by BCM, compared to those given by prior art.
Classification and notation system Case retrieval
a b s t r a c t
Thanks to a wide and dynamic research community on short term production scheduling, a large number of modelling options and solving methods have been developed in the recent years both in chemical pro duction and manufacturing domains. This trend is expected to grow in the future as the number of pub lications is constantly increasing because of industrial interest in the current economic context. The frame of this work is the development of a decision support system to work out an assignment strategy between scheduling problems, mathematical modelling options and appropriate solving methods. The system must answer the question about which model and which solution method should be applied to solve a new scheduling problem in the most convenient way. The decision support system is to be built on the foundations of CaseBasedReasoning (CBR). CBR is based on a data base which encompasses pre viously successful experiences. The three major contributions of this paper are: (i) the proposition of an extended and a more exhaustive classification and notation scheme in order to obtain an efficient sched uling case representation (based on previous ones), (ii) a method for bibliographic analysis used to per form a deep study to fill the case base on the one hand, and to examine the topics the more or the less examined in the scheduling domain and their evolution over time on the other hand, and (iii) the prop osition of criteria to extract relevant past experiences during the retrieval step of the CBR. The capabilities of our decision support system are illustrated through a case study with typical constraints related to process engineering production in beer industry.
Keywords: Fault Detection and Isolation, Extended Kalman Filter, Dynamic Hybrid Simulation, Distance, CaseBasedReasoning
1. Introduction
Nowadays, of safety and performance reasons, monitoring and supervision have an important role in process control. The complexity and the size of industrial systems induce an increasing number of process variables and make difficult the work of operators. In this context, a computer decision support tool seems to be wise. Nevertheless, the implementation of fault detection and diagnosis for stochastic system remains a challenging task. Various methods have been proposed in different industrial contexts [1]. They are generally classified as:
Due to both the limitations of traditional methods and the mutation of the industrial context, there is a need to find new efficient approaches to capitalize the new implicit and explicit design knowledge. As a consequence, different KBS have emerged in process design based on methods such as Conflict Based Approaches and CaseBasedReasoning (CBR). The first ones are based on modified TRIZ methods and tools to make them more easily applicable in the process engineering domain like in the studies of Li et al. (2003) and Negny et al. (2012) . These approaches are more focused on the phase of the research of new concepts. CBR is also suitable because numerous design problems become recurrent and these experiences can be easily reused. Their applications to assist in design decisions have been studied and improved for process design in the last decades as demonstrated in Negny et al., (2010) . But CBR suffers from three major draw- backs. The first two are knowledge elicitation and case adaptation. These drawbacks are commonly encountered in numerous CBR systems as proved by Chebel-Morello et al. (2013) , who explained that the time of knowledge workers dedicated to these phases is, respectively, 37.7% and 45.9% of their total time. The third draw- back is more specific to the application of CBR in design, where two categories of knowledge, i.e. contextual (corresponding to past experiences) and general (corresponding to rules, constraints, etc. ), must be combined to support a wide range of design decisions on the one hand, and to improve the quality of the solution on the other. Unfortunately, CBR systems only aim to encompass con- textual knowledge. Thus the challenge of this work is to raise the
Coudert, Thierry and Vareilles, Elise and Geneste, Laurent and Aldanondo, Michel Improvement of retrieval in Case-BasedReasoning for
system design. (2012) In: IEEE Industrial engineering and engineering
management - IEEM, 10-13 Dec 2012, Hong Kong, China.
Coudert, Thierry and Vareilles, Elise and Geneste, Laurent and Aldanondo, Michel Improvement of retrieval in Case-BasedReasoning for
system design. (2012) In: IEEE Industrial engineering and engineering
management - IEEM, 10-13 Dec 2012, Hong Kong, China.
Classification and notation system Case retrieval
a b s t r a c t
Thanks to a wide and dynamic research community on short term production scheduling, a large number of modelling options and solving methods have been developed in the recent years both in chemical pro duction and manufacturing domains. This trend is expected to grow in the future as the number of pub lications is constantly increasing because of industrial interest in the current economic context. The frame of this work is the development of a decision support system to work out an assignment strategy between scheduling problems, mathematical modelling options and appropriate solving methods. The system must answer the question about which model and which solution method should be applied to solve a new scheduling problem in the most convenient way. The decision support system is to be built on the foundations of CaseBasedReasoning (CBR). CBR is based on a data base which encompasses pre viously successful experiences. The three major contributions of this paper are: (i) the proposition of an extended and a more exhaustive classification and notation scheme in order to obtain an efficient sched uling case representation (based on previous ones), (ii) a method for bibliographic analysis used to per form a deep study to fill the case base on the one hand, and to examine the topics the more or the less examined in the scheduling domain and their evolution over time on the other hand, and (iii) the prop osition of criteria to extract relevant past experiences during the retrieval step of the CBR. The capabilities of our decision support system are illustrated through a case study with typical constraints related to process engineering production in beer industry.
3.1.3.5. Retain. It marks a significant stage in the informa- tion management cycle of case-basedreasoning. This is contributing to support the reuse of the recorded cases with their associated reasoning (establishment of diagnosis or search for treatment) for information retrieval, knowl- edge and information sharing and decision making. The structuration of cases base is determined by the indexing functions and the models of memory organization (simple model, model with dynamic memory or model based on categories) of the case-basedreasoning system. This struc- turation, for example, can use a network of categories and cases to explain organizational patterns according to the characteristics described by a name, a value and a level of importance regarding the membership of a case in a category ( Table 5 ).
McGill University, Montreal, Canada
Abstract
Bridge management systems (BMSs) are developed to assist decision-makers in optimizing the allocation of their limited budget on maintenance needs of bridge networks. Reliable deterioration models are essential constituents of BMSs that are used to predict the remaining service life of bridge components. The deterioration models incorporated in the recent BMSs have limitations that can be accepted for the analysis at the network level but not at the component level. Moreover, the current mechanistic deterioration models developed for the component level analysis are neither versatile nor adequately extensible to predict the service life of a large number of bridge components, or to incorporate additional deterioration parameters. Therefore, an artificial intelligence approach “Case-BasedReasoning (CBR)” is proposed to develop extensible, reliable, and generic deterioration models for the analysis at the component level. The CBR approach is utilized to predict the time to corrosion initiation of the reinforcing steel in concrete bridge decks. Data obtained from the Dickson Bridge in Montreal are used to generate the cases that populate the case library of the CBR model by applying the Monte Carlo Simulation techniques. Parameters that significantly affect the deterioration rate of concrete decks, such as the concrete cover thickness, apparent diffusion coefficient, and surface chloride concentration, are considered.
Among the most important Artificial Intelligence approaches expert systems have several drawbacks. The first of them is the time consuming aspect of the knowledge acquisition task especially in cases where few generic knowledge seem to exist. Besides, the binary scheme of the rules is not suited to the knowledge developed in scheduling problem. Upon the complexity of problem, simple rule-based systems do not seem to be efficient enough. The complexity of the decision process makes it difficult to construct a sufficiently complex neural network to model the resolution strategy too. Constraint propagation techniques and multi-agent systems are more applicable when the goal is to find a solution to a concrete problem rather than to look for a good resolution strategy. On the other hand, casebasedreasoning has numerous advantages. The reasoning can be started with relatively few initial knowledge. It is flexible and reactive, and the method is capable to learn in time which assures a continuous quality improvement. Therefore, among the possible candidates the Case-Based
Z. Lounis *
Institute for Research in Construction, National Research Council, Ottawa, Ontario, K1A 0R6, Canada
Abstract: This paper proposes a methodology for predicting the time to onset of corrosion of
reinforcing steel in concrete bridge decks while incorporating parameter uncertainty. It is based on the integration of artificial neural network (ANN), case-basedreasoning (CBR), mechanistic model, and Monte Carlo simulation (MCS). A probabilistic mechanistic model is used to generate the distribution of the time to corrosion initiation based on statistical models of the governing parameters obtained from field data. The proposed ANN and CBR models act as universal functional mapping tools to approximate the relationship between the input and output of the mechanistic model. These tools are integrated with the MCS technique to generate the distribution of the corrosion initiation time using the distributions of the governing parameters. The proposed methodology is applied to predict the time to corrosion initiation of the top reinforcing steel in the concrete deck of the Dickson Bridge in Montreal. This study demonstrates the feasibility, adequate reliability and computational efficiency of the proposed integrated ANN- MCS and CBR-MCS approaches for preliminary project–level and also network-level analyses.
This paper addresses the fulfillment of requirements related to case-basedreasoning (CBR) processes for system design. Considering that CBR processes are well suited for problem solving, the proposed method concerns the definition of an integrated CBR process in line with system engineering principles. After the definition of the requirements that the approach has to fulfill, an ontology is defined to capitalize knowledge about the design within concepts. Based on the ontology, mod- els are provided for requirements and solutions representation. Next, a recursive CBR process, suitable for system design, is provided. Uncertainty and designer preferences as well as ontological guidelines are considered during the requirements definition, the compatible cases retrieval, and the solution definition steps. This approach is designed to give flexibility within the CBR process as well as to provide guidelines to the designer. Such questions as the following are conjointly treated: how to guide the designer to be sure that the requirements are correctly defined and suitable for the retrieval step, how to retrieve cases when there are no available similarity measures, and how to enlarge the research scope during the retrieval step to obtain a sufficient panel of solutions. Finally, an example of system engineering in the aeronautic domain illustrates the proposed method. A testbed has been developed and carried out to evaluate the performance of the retrieval algorithm and a software prototype has been developed in order to test the approach. The outcome of this work is a recursive CBR process suitable to engineering design and compatible with standards. Requirements are modeled by means of flexible constraints, where the designer preferences are used to express the flexibility. Similar solutions can be retrieved even if similarity measures between features are not available. Simultaneously, ontological guidelines are used to guide the process and to aid the designer to express her/his preferences.
In this paper, a clinical decision support system (CDSS) [ 8 ] that relies on case-based reason- ing (CBR) is introduced, with the goal of helping practitioners to make decisions about the TAVI procedure.
The main concept of case-basedreasoning is to learn from previous experiences, even with a limited number of previous patient cases. This accumulated knowledge plays an essential role in decision making when facing new problems. The basic assumption of a CBR system is that similar cases should have similar solutions. CBR differs from other major artificial intelli- gence (AI) approaches, especially those that are based on learning process such as machine learning (ML), or other knowledge-based systems (e.g. rule-basedreasoning—RBR) [ 9 , 10 ]. CBR learns from previously processed cases, and the knowledge is progressively acquired [ 11 ]. The learning process is more evolutive than ML methods that require a special training phase, which is applied once from large datasets, to make future predictions. While CBR uses specific knowledge in the form of previous experience (the solved cases in the case-base), RBR, which is considered as pattern matching, represents general knowledge through a set of rules (if-then statements) [ 9 , 10 ]. The increased knowledge and experience in CBR becomes an advantage for medical applications when devices or clinical guidelines are continuously developed.
Due to both the limitations of traditional methods and the mutation of the industrial context, there is a need to find new efficient approaches to capitalize the new implicit and explicit design knowledge. As a consequence, different KBS have emerged in process design based on methods such as Conflict Based Approaches and CaseBasedReasoning (CBR). The first ones are based on modified TRIZ methods and tools to make them more easily applicable in the process engineering domain like in the studies of Li et al. (2003) and Negny et al. (2012) . These approaches are more focused on the phase of the research of new concepts. CBR is also suitable because numerous design problems become recurrent and these experiences can be easily reused. Their applications to assist in design decisions have been studied and improved for process design in the last decades as demonstrated in Negny et al., (2010) . But CBR suffers from three major draw- backs. The first two are knowledge elicitation and case adaptation. These drawbacks are commonly encountered in numerous CBR systems as proved by Chebel-Morello et al. (2013) , who explained that the time of knowledge workers dedicated to these phases is, respectively, 37.7% and 45.9% of their total time. The third draw- back is more specific to the application of CBR in design, where two categories of knowledge, i.e. contextual (corresponding to past experiences) and general (corresponding to rules, constraints, etc. ), must be combined to support a wide range of design decisions on the one hand, and to improve the quality of the solution on the other. Unfortunately, CBR systems only aim to encompass con- textual knowledge. Thus the challenge of this work is to raise the
example simulation and optimization. But with this approach, the possible actions are very limited and specific to the object to design. Another approach is to exploit the experiences gained during earlier design because they allow to reduce the delay of design since some choices are no longer to make or to question. In this context, some firms want to have meth- ods and tools to support design exploiting past knowledge. A design support system, needs the representation of the knowl- edge within a firm (or a profession) in order to exploit it and to facilitate the development of new objects. Various techniques coming from Artificial Intelligence has been developed to rep- resent, to capitalize and to exploit knowledge for the problem of support design. Case-basedreasoning (CBR) is one of them. In the whole chaining steps of the process design, CBR has been widely used (in every technical domains) as a decision support system. In the majority of cases, CBR systems are limited to products design where one or two tens of com- ponents interact. The CBR method is based on analogical reasoning inside a specific domain (technical or not). This method is a Knowledge Management one, used to capital- ize, to store and to reuse knowledge and earlier experiences. CBR has recently appeared in chemical engineering with applications in: process design by reusing flowsheets (Surma and Braunschweig, 1996), synthesis of process separation (Pajula et al., 2001), reactive distillation (Avramenko et al., 2004), mixing equipment selection (Kraslawski et al., 1995), minimisation of environmental impact (King et al., 1999), and generation of process alternatives (Lopez-Arevalo et al., 2007).
7 Conclusion
Case-basedreasoning systems use similarity (usually in the form of a similarity measure or a distance). This is obvious for the retrieval of a case similar to the target case but this chapter shows how it can be used for adaptation: an important class of revision operators is based on distances. Indeed, u d -based adaptation can be reformulated as the process of selecting the case instances that are the closest ones to the source case, in the metric space (U , d), with constraints given by DK. Since adaptation aims at solving a certain type of analogical problem (in which two of the four elements in the analogy are problems, the other ones—including the unknown—are solutions), this approach concretely relates analogical reasoning with belief revision.
a b s t r a c t
Telemedicine is the medical practice of information exchanged from one location to another through electronic communications to improve the delivery of health care services. This research article describes a telemedicine framework with knowledge engineering using taxonomic reasoning of ontology modeling and semantic similarity. In addition to being a precious support in the procedure of medical decision-making, this framework can be used to strengthen significant collaborations and traceability that are important for the development of official deployment of telemedicine applications. Adequate mechanisms for information management with traceability of the reasoning process are also essential in the fields of epidemiology and public health. In this paper we enrich the case-basedreasoning process by taking into account former evidence-based knowledge. We use the regular four steps approach and implement an additional (iii) step: (i) establish diagnosis, (ii) retrieve treatment, (iii) apply evidence, (iv) adaptation, (v) retain. Each step is performed using tools from knowledge engineering and information processing (natural language processing, ontology, indexation, algorithm, etc.). The case representation is done by the taxonomy component of a medical ontology model. The proposed approach is illustrated with an example from the oncology domain. Medical ontology allows a good and efficient modeling of the patient and his treatment. We are pointing up the role of evidences and specialist’s opinions in effectiveness and safety of care.
4 Interpolative ReasoningCase-basedreasoning relates two similar problems and transfers the solution of one of them to the other one. An analogical proportion states particular similarity and dissimilarity relations between f our terms. Thus, case-basedreasoning and ana- logical reasoning are two forms of similarity-basedreasoning. But they are not the only ones. In this last section of the chapter we present a brief overview of studies based on another similarity-basedreasoning: the interpolative (and extrapolative) reasoning. Interpolation allows us, when the current situation is intermediate be- tween known situations, to conclude in an intermediate way with respect to the conclusions of these situations. When the conclusion of only one situation, close to the current situation, is known, a solution can be extrapolated for the current situa- tion, provided that some available information about the variations around this close situation can be exploited. Therefore, interpolation and extrapolation need variables with ordered referentials and some notions of similarity. These forms of reasoning, though they are important in commonsense reasoning, have got very little attention in AI outside the community working on fuzzy sets and approximate reasoning. First, some recalls about fuzzy sets and approximate reasoning are given. Then, in- terpolation and extrapolation in this framework are discussed. Finally, some studies on this subject that are not based on fuzzy sets are briefly presented.
∗ Corresponding author. Tel.: +330534323663.
second question is how to transform a waste into new valuable products. To solve these questions, authors propose to use an arti- ficial intelligence system, and more particularly casebased reason- ing (CBR). CBR is relevant for this kind of problems because it al- lows solving problems without a clearly defined knowledge of the process needed for the resolution. The reasoning can rely on a vast number of cases, with their precise description of previous solved problems and their associated solutions ( Cordier, Mascret, Mille, 2009 ). Secondly, in the domain of waste treatment, cases may contain different information: valorisation processes and essences for the new created objects. In the literature, casebased reason- ing systems are used in different waste treatment problems and in processes research. For example, López-Arévalo, Bañares Alcán- tara, Aldea, Rodríguez-Martínez, and Jiménez (2007) describe a tool based on CBR for the generation of process alternatives. Yang and Chen (2011) propose a classical CBR retrieve method used for Eco- innovation Kuo (2010) gives an example of CBR used to determine a recyclable index of some components. Liu and Yu (2009) use CBR for problems linked to environmental topic. Zeid, M. Gupta, and Bardasz (1997) propose a model dedicated to disassembling problems.