Intelligent evaluation approach for electronic product recycling via case-based reasoning

https://doi.org/10.1016/j.aei.2005.11.003Get rights and content

Abstract

Due to worldwide stricter regulations on end-of-life product recycling, recyclability of more consumer products would be of concern. Evaluation of recycling cost and benefit of reclaimed materials as well as selection of recycling strategy is essential for not only product designers but recycling policy makers. This study proposes an intelligent evaluation approach that incorporates case-based reasoning models, economic analysis model and domain expertise. Case-based reasoning method is used to determine recycling strategy and performance of disassembly operation while economic analysis model is used for quick estimating recycling cost and benefit. Domain expertise is adopted in model parameter estimation to facilitate the computation. The approach can be used both as a module of design for recycling in eco-design evaluation and as an independent evaluation tool for recycling policy making. Illustrative examples are included to depict the application potential. Advantages and drawback are discussed.

Introduction

Due to stricter product recycling regulations, such as WEEE, end-of-life (EOL) product recycling evaluation has become an important issue. Local government puts more efforts in policy making of product recycling. For example, in Taiwan, policy making officers are interested in which product should be included in national mandatory recycling policy and how much it would cost for recycling the specific product and whether subsidization policy is needed. On the other hand, from product designers' point of view, a module of ‘design for recycling (DfR)’ becomes inevitable in eco-design process. This study tries to provide a simple and quick solution for the questions of how to determine recycling strategy and how to get a quick estimation of recycling cost and benefit.

Several studies have focused on recycling strategies selection. Rose [1] proposed a system called ELDA (end-of-life design adviser) to help determine recycling policy and conducted a survey to verify their work. She chose six product characteristics including wear-out life, technology cycle, level of integration, number of parts, reason for redesign and design cycle and constructed a decision tree model to find a best recycling strategy. Zhang et al. [2] adopted analytical hierarchic process (AHP) to find best recycling strategy. The AHP based evaluation considered environmental impact, cost and reclaimed materials as the major criteria for strategy determination. Bras et al. [3] proposed recycling and remanufacturing indices to define the product recyclability. These indices can be used as performance indicators in selecting best recycling strategy. Other examples discuss recycling strategy selection include Krikke et al. [4], Linton [5] and Wu [6].

On the other hand, many literatures present result on recycling cost estimation. Most of them took bottom-up approach where the estimation is conducted based on operation breakdown and summation of detail cost items. For example, Boothroyd and Atling [7], Gungor and Gupta [8], Zhang and Kuo [9], Feldman [10], Ichikawa [11] systematically investigated the disassembly sequence and related operations so that disassembly cost can be accurately estimated. For material recovery part, several studies estimate recycling cost including disassembly cost, shredding and separation cost, disposition cost and revenue from reclaimed materials. Examples include Tsao [12], Ichikawa [11], and Stutz [13]. Some studies have incorporated both cost estimation and environmental impact estimation. For example, Lee et al. [14] tried to find alternative that can both maximizes profit and minimizes environmental impact and took coffee maker as an example. Hula et al. [15] had a similar approach and solved the optimization problem via genetic algorithm method.

Case-based reasoning (CBR) was first proposed by Schank and Abelson [16] as one of the artificial intelligence techniques. CBR is different from the traditional rule-based reasoning methods where past experiences (cases) are stored in a case library to provide useful information for new problem (case). Kolodner [17] suggested six-step procedures for CBR: (1) retrieve cases, (2) find ballpark solution with indices, (3) justify the solution and sometimes adapt the cases, (4) criticize or revise the case base, (5) evaluate the inference results, and finally (6) store the case. Goodman [18] summarized that CBR has advantages when: (1) domain knowledge is difficult to be formalized, (2) solid cases library exists, (3) there are too many rules if rule-based reasoning is used, (4) domain is dynamic while quick response and inference are required, and (5) goal is to find successful experience as the solution base.

In this study, CBR method is adopted to (1) select appropriate EOL strategy and (2) estimate disassembly time and the weight percentage of parts being disassembled before shredding process. As the successful EOL cases appear in many areas in the world, one can extract experience from past cases when including a new product in EOL recycling and treatment system. A case-based approach (stage I) is proposed in this study comparing to the rule-based approach in the literature such as Rose [1].

On the other hand, former literature such as Boothroyd and Atling [7] and Ichikawa [11] estimate disassembly time and cost requiring detail information such as disassembly motion and time information, parts' geometry and disassembly sequence. The break point between disassembly and shredding for material recovery has to be determined before further cost estimation. The conventional methods need detail information and trade-off analysis. This study proposes a CBR approach (stage II) to provide relatively rough but quick estimates of disassembly time and break point between disassembly and shredding (i.e. weight percentage of parts disassembled) for disassembly cost estimation. Based on the CBR results, the cost and benefit estimation can be done separately for disassembly and processes after shredding. As the number of types of products being successfully recycled worldwide increases, this CBR approach could be used to get a disassembly cost estimate for evaluating new EOL products without doing detail experiment and data acquisition. Please note that the two CBR models (stages I and II) can be applied independently.

In Section 2, an overall picture of the proposed approach is presented. Section 3 discusses the CBR based model for determining best recycling strategy. Illustrative examples are presented comparing the results with previous work. There are two subsections in Section 4. The first subsection presents another CBR model for estimating disassembly time and the material portion taken by disassembly. The second subsection presents a simplified economic evaluation model for cost–benefit estimation. Domain expertise is adopted in model parameter determination to facilitate the computation. Section 5 presents the application potential of the proposed cost–benefit evaluation method via four example EOL products. At the end, conclusions are presented.

Section snippets

Overview of the proposed CBR based approach

The proposed approach facilitates EOL recycling strategy selection and cost–benefit estimation in two stages. Different CBR models are used in the two stages. Fig. 1 presents the flowchart of the proposed CBR based evaluation process.

  • Stage I:

    In the first stage, the CBR based model is used to determine the best EOL strategy. Fifty cases from Rose [1] database are adopted in the case library where case indices are defined based on the decision variables in Rose [1] work. Candidate recycling strategies

CBR model for EOL strategy determination

The essence of case-based reasoning (CBR) is to find very ‘similar’ past cases and extract the experiences from that cases for the new problem. In this study, existing successful cases of product recycling are adopted from Rose [1] for extracting recycling experience. Most of the cases are electronic and electrical products while few of them are mechanical products. Case indices including wear-out life, technology cycle, level of integration, numbers of parts, reason for redesign and design

Cost–benefit estimation for product recycling

In the second stage of the proposed approach, an economic evaluation model is proposed to get a quick estimation of recycling cost and revenue. The estimation outputs include disassembly cost, recycling cost, disposal cost, treatment cost for hazardous materials and the revenue from selling reclaimed materials. In the followings, two subsections discuss the proposed methods for disassembly and material recovery phases, respectively.

Illustrative examples

In this section, four EOL products are illustrated using the proposed approach. Examples include cellular phone, commercial-use refrigerator, LCD monitor and telephone set. Reason of picking them is because local government is considering inclusion of them in national mandatory regulations. The upper half of Table 4 shows the results of the CBR model I proposed in Section 3. The most similar cases are depicted in the second row while the similarity score and the recommended EOL strategies are

Conclusions

This study proposes CBR based methods to determine product EOL strategy and conduct cost-benefit estimation. First, a CBR I model is proposed to find the most similar past cases and therefore select an appropriate EOL strategy. It provides an alternative method for selecting EOL strategy other than the conventional rule-based methods. Secondly, CBR II model is built to estimate disassembly time and weight percentage of parts being disassembled for further cost-benefit analysis. The two

Acknowledgements

This work was supported by the National Science Council of the Republic of China (NSC-93-2621-Z-006-010). Authors would also like to thank three anonymous reviewers and Dr Yasushi Umeda for the thoughtful suggestions that helped improve the readability of the manuscript.

References (23)

  • K. Feldmann et al.

    Computer-based product structure analysis for technical goods regarding optimal end-of-life strategies

    Eng Manufact

    (2001)
  • Cited by (40)

    • Disassembly

      2023, Design and Manufacture of Structural Composites
    • A design tool to diagnose product recyclability during product design phase

      2017, Journal of Cleaner Production
      Citation Excerpt :

      There are occasions when disassembling the whole product into its smaller parts is impracticable due to the presence of hazardous materials (Desai and Mital, 2003). The presence of hazardous materials generates higher recycling costs due to the need for treatment and disposal of these materials (Shih et al., 2006). Each hazardous material has its own procedure for treatment and disposal to avoid damaging human health and the environment.

    • Ontology and CBR based automated decision-making method for the disassembly of mechanical products

      2016, Advanced Engineering Informatics
      Citation Excerpt :

      In this paper, ontology and CBR are applied to the automated decision-making for the disassembly of mechanical products. Compared with other related methods [9–14], the proposed method has following advantages: High flexibility and expandability.

    View all citing articles on Scopus
    View full text