Development of a justification tool for advanced manufacturing technologies: system-wide benefits value analysis

J. Eng. Technol. Manage. 18 (2001) 157–184

Development of a justification tool for advanced manufacturing technologies: system-wide benefits value analysis Sharon M. Ordoobadi∗ , Nancy J. Mulvaney IMSE Department, College of Engineering, 237 Durland Hall, Kansas State University, Manhattan, KS 66506, USA

Abstract Growing competition and increasing demands from customers are forcing small manufacturers to consider investments in advanced manufacturing technologies (AMTs). For many reasons, such investments are often difficult to justify by means of a traditional economic analysis alone. As a result, it is often necessary to consider the system wide benefits associated with AMTs in order to justify their adoption. A process known as system wide benefits value analysis (SWBVA) has been developed to assist decision makers with their advanced technology decisions. Users of the tool first perform an economic analysis to see if the investment is economically justified. If it is not yet justified, the gap between the minimum desired economic return and the actual return amount is calculated. Users can follow a series of procedures to determine if the value of the system wide benefits associated with the advanced technology is sufficient enough to justify this gap. These procedures involve customizing a formal model of system wide benefits to suit the technology decision being evaluated, setting desired goals for each benefit being considered, and answering a series of input questions about the level of those benefits they feel can be obtained from such a technology. A fuzzy expert system is the internal mechanism used to manipulate user inputs into crisp output values for each benefit category. If the determined output values for each system wide benefit are greater than or equal to the user-defined benefit goals, then the gap amount is believed to be justified. Users are provided with a summary report on the calculated results and are allowed to readjust their benefit goals and repeat the analysis if necessary. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Advanced manufacturing technology; System-wide benefits; Justification tool; Fuzzy expert system

∗ Corresponding author. Tel.: +1-785-532-5606; fax: +1-785-532-7810. E-mail address: [email protected] (S.M. Ordoobadi).

0923-4748/01/$ – see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 3 - 4 7 4 8 ( 0 1 ) 0 0 0 3 3 - 9

158

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

1. Introduction Innovation and the adoption of advanced technologies are crucial activities for manufacturing firms today. This is especially true for smaller manufacturing firms which have not necessarily established formal processes for research and development (R&D). R&D challenges facing small manufacturers are numerous (Wallsten, 1998). Small manufacturers often take on an antiquated philosophy when it comes to innovation, and although there seems to be no scarcity of good ideas within these companies, the ability to take ideas and develop them into new products or processes is often lacking. The evaluation tool proposed here is but one aspect of a larger project that is underway on the management of technological innovations (MOTI) for small manufacturers. The main purpose of this paper is to identify and analyze the technological, psychological, and behavioral barriers that inhibit the innovation process in smaller firms, and to likewise uncover the success factors that make prosperous innovations possible. Through preliminary interviews with small manufacturers in Kansas, it was determined that this objective could best be met through a tool that provided examples of past successes and failures in the innovation process, while serving as a useful guideline for gaining insight and making decisions about current innovative ideas. This main aspect of the project is partially fulfilled through the collection and organization of a database of cases that describe examples of past manufacturing innovation successes and failures. The case database has been developed through a series of interviews and focus group discussions with chief executive officers (CEOs), presidents, and managers of small manufacturing firms in Kansas. Two tools are being developed simultaneously to utilize the case database and fulfill the main objectives of the project: (1) a tool which utilizes case grammar to classify past cases and draw comparisons to current situations and (2) a tool which applies analogical reasoning and proverbs to describe past cases and make comparisons to current scenarios (Xue, 1999). The evaluation tool presented here is meant to be a supplementary tool to this larger aspect of the project. It is felt that once firms have the opportunity to analyze their current ideas through the examination of cases in the aforementioned tools, a mechanism for formally evaluating those projects then becomes necessary. The methodology presented here will enable decision makers in small manufacturing firms to make informed and complete analyses of their potential advanced technology projects. Such analyses will consider not only the traditional cost factors involved in such investments, but also the secondary, system wide benefits that can often be obtained with advanced technologies. The following sections will describe the proposed tool, system wide benefits value analysis (SWBVA), in further detail. First, a comprehensive outline of the problems associated with evaluating advanced technologies along with an analysis of the literature on past evaluation methods is presented. The next section includes an introduction to the SWBVA along with a depiction of the benefits model and a description of the SWBVA process. A synopsis of the data collection modes for the modeling aspects of the system is provided next, followed by an account of the modeling procedures themselves. Finally, an example to illustrate application of the tool is offered along with some conclusions and suggestions for future research.

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

159

2. Literature review 2.1. Problem identification The economic justification of advanced manufacturing technologies (AMTs) has been a topic of question and concern for quite some time. AMTs can be defined as any type of advanced technology that, when incorporated into a manufacturing operation, has a significant impact on the product, process, and informational aspects of the system. Some even consider an AMT to be a business strategy in and of itself (Wilkes and Samuels, 1991). More specifically, AMT investments typically include (but are not limited to) the following (Small and Chen, 1995) systems. • Stand-alone systems, such as computer-aided design (CAD). • Intermediate systems, such as automatic storage and retrieval systems (AS/RS). • Integrated systems, such as flexible manufacturing cells (FMC/FMS). In addition, a common agreement is that AMT investments have several discernable characteristics (Wilkes and Samuels, 1991). • • • •

A long life and attunement. Investment requirements that expand to several years. Increasing returns over time. Various intangible, or system wide benefits.

It is these characteristics that contribute to the major problems in evaluating and justifying advanced technologies. First, AMTs are often very difficult to justify on the basis of strict economic evaluations alone (i.e. net present value, NPV). The reasons behind this problem can be summarized as follows (Canada and Sullivan, 1990). • These types of technologies often require extremely high initial capital costs that are not easily justified through traditional methods. • Many companies impose extremely high hurdle rates. In fact, many firms require minimum attractive rates of return (MARR) of 30%, and/or paybacks of less than 2–3 years for AMT investments. • Companies have the tendency to manipulate their MARRs, paybacks, etc. specifically for AMT investments because of the high levels of risk associated with them. Based on economic means alone, such requirements make it nearly impossible for AMT projects to be justified. Instead, only those projects that can show a big profit early on will be considered legitimate. • Firms tend to compare investments in advanced technologies with the status quo. This means that companies will consider a ‘do nothing’ approach as a valid option when pursuing their investment alternatives. However, companies that compare an AMT investment option with the status quo are refusing to realistically view the risks and opportunity costs associated with this decision. The difficulty associated with traditional economic models and their inability to properly justify AMT investments leads to another problem area. Many of the benefits associated with these technologies are often non-quantifiable in nature and are difficult to estimate.

160

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

Moreover, these benefits have multiple effects that reach every aspect of the manufacturing system. This deeper categorization of benefits will be referred to as a system wide benefit, and is formally defined as any benefit that adds value to the manufacturing system, but is not necessarily measurable (or tangible in a dollar value) for use in any traditional economic analysis technique. System wide benefits tend to add great significance to a manufacturing system, yet managers, accountants, and other decision makers are often not even aware of their existence because of the lack of a formal outline and explanation of their meaning and impact. Often, it is these types of benefits that will “push a project over the edge” of the justification hurdle and make investments in AMTs visibly favorable. Although there are many techniques currently available for considering such benefits, no current method formally takes into account a full list of system wide benefits. In addition, many of the current models simply attempt to rate or blindly estimate the value of intangible benefits, with no suggestion for determining any real level of contribution that the benefits make to the manufacturing system. The problem, therefore, lies in the challenge of officially incorporating the system wide benefits into the evaluation procedures, so that the full recognition of benefits can logically and accurately be taken into consideration. 2.2. Analysis of evaluation techniques Fig. 1 provides a summary of past techniques used in AMT investment decisions (adapted from Badiru et al., 1991 and Meredith and Suresh, 1986). Many specific applications of these techniques were found in the literature. A detailed description of each will not be provided here, but Table 1 provides a categorization of the literature for each of the applications found as they pertain to the techniques listed in Fig. 1. For a comprehensive overview of each of the models listed in Table 1, refer to Mulvaney (1998).

Fig. 1. Common evaluation techniques for AMT investment decisions.

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

161

Table 1 Categorization of references for AMT evaluation models and techniques Title of model or technique

Author(s)

Selection of a technique based upon level of integration

Meredith and Hill (1987) Meredith and Suresh (1986)

Simple estimation method

Azzone and Bertele (1991) Bennett and Hendricks (1987) Downing (1989) Noaker (1994) Noble (1990) Pant and Ruff (1995) Swann and O’Keefe (1990)

Breakeven approach (x-gap approach)

Finnie (1988) Kaplan (1986) Wilkes and Samuels (1991)

Systems value analysis

Badiru et al. (1991) Troxler and Blank (1989)

Scoring models

Dhavale (1995) Parsaei et al. (1988) Parsaei and Wilhelm (1989) Soni et al. (1990) Sullivan (1986)

Analytic hierarchy process (AHP)

Albayrakoglu (1996) Datta et al. (1992) MacStravic and Boucher (1992) Mohanty and Venkataraman (1993) O’Brien and Smith (1993) Oeltjenbruns et al. (1995) Putrus (1991) Weber (1993)

Simulation

Kassicieh et al. (1993) Son (1993)

Strategic methods

Downing (1989) Kakati and Dhar (1991)

Other unique methods

Presley et al. (1995) Son and Park (1987) Swamidass (1987)

Several problems associated with these techniques are outlined below. • The accuracy of some of the models is questionable Although no method can be expected to produce 100% accuracy, some of the techniques are often lacking in precision. For instance, the analytic hierarchy process (AHP) and scoring methods take into account the ‘importance’ of the benefits with regard to a certain goal or alternative, while neglecting to highlight a specific contribution level for each benefit. Likewise, the breakeven approach often recommends that a total benefit

162

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

value be considered as potentially obtainable rather than analyzing the contribution of each individual system wide benefit in turn. The simple estimation method is lacking in precision, because it requires that decision makers make a ballpark guess as to the dollar value of certain benefits. • Some of the methods are extremely time-consuming This problem lies especially in models such as the AHP, systems value analysis (where AHP is the core evaluating tool), and the various unique models. Pairwise comparisons which are made in the AHP process often take a tremendous amount of time for the decision maker, especially when consistency checks are required. As the number of criteria and alternatives increases, the time involved with these methods expands dramatically. In addition, many of the unique methods require extreme amounts of data that may or may not be readily available to decision makers. • Some of the models are too simple For instance, a strategic assessment could be very useful for evaluating the long-term impacts of AMT investments, however, just because an alternative is strategically appealing does not necessarily mean that the benefits associated with that alternative are guaranteed. Some other type of analysis is needed to make this a more complete method for analyzing AMT alternatives. In the same way, many of the methods do not fully take into account all of the potential system wide benefits that can be obtained with AMTs. No one method or model found in the literature provided the decision- maker with a complete, formal model of the system wide advantages available with the acquisition of AMTs. Many times, the models would make a few suggestions and then leave the generation of the benefits up to the decision-maker. This leaves no formal method for educating or reminding the decision-maker of what types of benefits exist with these technologies. • Some of the models are too specific Some of the methods are very individualistic in nature and, therefore, cannot be generalized across companies. Furthermore, these methods require substantial commitment of company time and other resources. In summary, the currently available AMT justification techniques do not lend themselves very well to the peculiar needs of smaller firms. Small manufacturers need justification methods that are easy to understand, affordable, easy to use, and not very time consuming. Most available justification techniques, however, are expensive, time consuming, and difficult to use due to high level of sophistication.

3. Research design In previous section we addressed the problems with applying currently available justification methods and models to the peculiar needs of smaller manufacturers. In recognition of these shortcomings, the objective of this research is to develop a justification tool that will assist decision makers of small manufacturing firms in making more informed and complete analyses of their AMT proposals. This is realized through a two-step process, outlined as follows.

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

163

1. Establishment of a formal model of system wide benefits for consideration in AMT investment decisions. 2. Design of a methodology for evaluating AMTs, which will utilize the formal model of system wide benefits and facilitate decision makers in the determination of a value for the benefits associated with a particular technology. 3.1. Introduction to the SWBVA System wide benefits value analysis actually utilizes some of the methods from the literature, but attempts to combine them in a way that helps to overcome many of their perceived problems. Specifically, the method of systems value analysis is used to combine the system wide benefits documented from the literature into a formal and complete model. The breakeven or x-gap approach for identifying the amount of system wide benefits needed to justify the project is also used. The difference in the proposed tool starts at this point. A fuzzy expert system is the core tool used to model the benefits into determinable membership functions and therefore allow decision-makers to enter specific estimates about each project based upon the membership values. The proposed tool is believed to overcome the problems associated with other methods by providing a formal and detailed model of system wide benefits, and by defining a level of contribution for each benefit type, while requiring a reasonably low amount of time and input from the evaluator. These provisions will help to ease the problems of accuracy, simplicity, and time mentioned in the previous section. In addition, the design of the proposed tool allows for the model to be easily adapted and does not require complicated implementation efforts within each company. This helps to overcome the problem of specificity. 3.2. The system wide benefits model Various system wide benefits attributable to the adoption of advanced technologies were uncovered in the literature. Upon analyzing this list of benefits, several inter-relationships and levels of abstraction were discovered. For the purposes of the SWBVA, it was felt that there was a need to systematically form the benefits into a categorized model that would realistically depict the varying degrees of specificity and the corresponding relationships that existed. In doing so, three hierarchical levels were uncovered: (1) benefits; (2) sub-benefits and (3) benefit indicators. The process of forming the benefits into a hierarchical model was performed by logically forming all benefits found in the literature into these three categories from the most abstract to the least abstract based upon their inter-relationships with each other. For instance, it is logical to group increased product flexibility as a sub-benefit of increased flexibility, since it is, in fact, one component of the flexibility factor. The formalized model of system wide benefits can be seen in Fig. 2. The model of system wide benefits follows the systems value analysis method for organizing and analyzing deterministic system value. The development of a systems value analysis model requires the completion of three main activities: (1) establishment of the various components that characterize the system; (2) manipulation of the components into logical levels and sub-levels and (3) combination of the levels and the measurement of their contribution of value to the system (Troxler and Blank, 1989). The development of Fig. 2

164

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

Fig. 2. The SWBVA benefits model.

exemplifies the first two steps of this process in the following way: (1) the identification of the benefits through the literature represents the establishment of the components of manufacturing system value and (2) the organization of the benefits into a three-tiered hierarchy represents the division of the components into logical levels and sub-levels. Not all of the benefit indicators have been listed in Fig. 2, but a sampling of the types of indicators to be considered for each benefit category has been included. Theoretically, the value that is added to a manufacturing system through the consideration of system wide benefits resulting from the adoption of AMTs can be measured by working up the level of abstraction (from the least abstract level to the most abstract) in the hierarchy to determine a final level of worth. For example, one of the benefit categories has been defined as increased flexibility. A sub-benefit of increased flexibility is product flexibility, which in turn possesses several indicators including the Number of product types that can be manufactured with the new technology and shorter cycle times and set-ups. From this, it can be concluded that if the adoption of an advanced technology prompts an increase in the number of product types that can be manufactured and/or a decrease in cycle times

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

165

and set-ups, then an increase in product flexibility is likely. This likewise contributes to an overall amount of increased flexibility that is produced within the manufacturing system. The manipulation of all such indicators, sub-benefits, and benefits throughout the model will demonstrate the level of system wide benefit value that is added to the manufacturing system through the adoption of advanced technologies. 3.3. The SWBVA process Now that a formal model of system wide benefits has been introduced and the foundation of the system explained, it is necessary to outline the details of the SWBVA process. Fig. 3 depicts a process flow diagram for the SWBVA. The process of the SWBVA consists of

Fig. 3. Process flow diagram for SWBVA.

166

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

10 main steps, each of which is further explained below. An asterisk (∗ ) preceding the description of a step indicates that this particular step is an internal function performed solely within the system and requires no input or action from the decision-maker. 1. Perform an economic analysis of the proposed AMT using system wide benefits model Once an advanced technology is proposed for adoption, the first step is to perform an economic analysis using a traditional evaluation technique. For the purposes of the SWBVA, the internal rate of return (IRR), NPV, and payback period methods will be provided to carry out this step. The expanded economic analysis incorporates all of the readily measurable costs and benefits of the proposed technology as well as measurable benefits that are included in the system wide benefits model. Thus, making this economic analysis more complete and thorough compare to traditional evaluation techniques. 2. ∗ Narrow and sort the model of system wide benefits To avoid double counting of benefits, the system wide benefits that are measurable and included in the financial analysis of step 1 will be eliminated from the model of system wide benefits. 3. Examine the results to see if the project is justified Once the economic analysis is completed, the results should be compared to decision criteria (i.e. the IRR greater than or equal to minimum attractive rate of return (MARR), the NPV ≥ 0, or the payback period less than or equal to target lend) to see if the project is justified through this analysis. If the project is in fact justified, then no further analysis is necessary. 4. ∗ Calculate the x-gap If the project is not justified through the economic analysis of step 1, then the difference between the result of the economic analysis and the minimum required justification level is calculated. This value is hereafter referred to as the x-gap. A definition of the x-gap will likewise be explained to the user of the system. 5. Examine the x-gap to determine further action The value of the x-gap is analyzed to determine, whether continuation of the SWBVA process is warranted. In some instances where the x-gap is extremely small, the decisionmaker may feel that the project is close enough to the minimum required justification level and therefore no further analysis is necessary. In other instances, the decision-maker may feel that the x-gap is so large that no amount of system wide benefits will justify the proposed technology. The decision regarding the value of the x-gap and the necessity to analyze the project further will be left up to the decision maker, who will simply be prompted as to whether they wish to continue. Factors such as industry, firm’s size, and type and cost of the advanced technology under consideration can affect the decision-maker’s acceptable level of x-gap. In addition, prior investment decisions taken by others regarding similar technology can be used as guidelines to determine the level of x-gap that would warrant further analysis. 6. Answer initial calibration questions The decision-maker will be required to answer a series of questions about the utility of the various benefit categories to the particular AMT investment being considered. The user is provided with the full list of benefits (Fig. 2), excluding the ones that are accounted for in the economic analysis of step 1. Only those benefits that are viewed as

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

7.

8.

9.

10.

167

useful for consideration to the project at hand will be evaluated in the SWBVA. At this time, the decision maker will also set benefit goals for each benefit category considered. These benefit goals are used as the decision criteria in the SWBVA evaluation and will be compared to the final outputs of the fuzzy procedures to determine whether justification of the x-gap has been obtained. ∗ Narrow and sort the model of system wide benefits Those benefits that are not considered useful for inclusion in the SWBVA will be eliminated from the model. Complete fuzzy input procedures The decision maker will be required to answer a series of questions which will draw on the established fuzzy membership functions to determine values for the SWBVA components. Refer to Mulvaney (1998), for a complete definition and description of the fuzzy membership function. ∗ Monitor the results to determine when and if the benefit goals are obtained As the decision maker answers the fuzzy input questions, the system will keep a running account of the SWBVA to determine when and if the benefit goals are achieved. If all the benefit goals are reached during the first analysis attempt, the user will be notified of this and a summary report will be generated. Examine the summary report to evaluate further actions Once the system wide benefit fuzzy input questions have been answered and crisp output values have been internally determined, the user will then be informed as to whether or not their benefit goals have been reached. A summary report will be presented to the user which details whether each benefit goal has been met. If all benefit goals have been reached, then the evaluation is complete and the user is informed that the amount of system wide benefits associated with their particular investment decision satisfies their desired benefit amounts for justifying the x-gap. If all of the benefit goals have not been met, the summary report will indicate to the user which benefits are lacking and will be given the option to readjust the goals to see for what values the x-gap is fulfilled. If the decision-maker is willing to readjust the benefit goals, the analysis is then repeated (steps 6–9). In addition to the individual sub-benefit results, averages for each benefit category and an overall average among all benefit categories (one overall value for all system wide benefits) will be presented in the summary report. Although the main basis of the analysis is driven from the sub-benefit and indicator categories, these overall values may provide the user with some additional information from which to base their final investment decision. For example, even though the benefit goal for increased demand flexibility may not be met, if the overall value for increased flexibility is high enough, the decision maker may decide that this is sufficient. The presentation of a complete picture of the value of system wide benefits is meant to provide the user with the greatest amount of information possible from which to base their decision.

4. Data collection modes After completing the initial economic analysis and calibration procedures for SWBVA, the remaining steps of the process involve completing the fuzzy input procedures to arrive at

168

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

crisp output values for comparison with the benefit goals. However, before these processes can be described further and the methodologies behind their development explained, it is necessary to establish the data basis for these procedures. Two surveys were conducted to establish a data basis for the SWBVA. Survey targets were CEOs, Presidents, and Managers of small manufacturing firms in Kansas. These experts were selected as survey targets because of their background and experience in owning and operating small manufacturing firms. All firms were small manufacturing companies that were current or potential clients of manufacturing learning center, a technology transfer center in Kansas, which to a certain degree implies their interest and experience in AMT. The majority of the responding firms had over 25 years of operation, from 50 to over 150 employees, were private organizations, and serviced primarily industrial markets. Survey 1 was conducted in order to extract the expert judgment of the decision makers to refine the model of system wide benefits and include only those indicators that were the greatest designators for each sub-benefit. That is, to determine what indicators are best for consideration in the SWBVA tool. Utilizing all indicators that can be generated from the literature or other expert sources would make the process incredibly time consuming for a decision maker attempting to use the tool. In addition, some indicators may be better representatives than others for a certain sub-benefit category. To resolve these issues, it was decided that a survey of expert decision makers like those that would be using the SWBVA tool in the future would be the best source for determining what indicators should be considered. If the model of system-wide benefits could be refined by these experts to include only those indicators that were the greatest designators for each sub-benefit, then the amount of time and effort required by the decision maker in using the SWBVA tool would be greatly decreased. Survey 1 first asked a series of company profile questions, including the number of employees, years of operation, organization type, and market type. The remainder of the survey contained a table which listed each sub-benefit category and its definition. Respondents were then asked to rank a list of indicators in the order of their impact to that particular sub-benefit, with a rank of one suggesting the best indicator. Once the model of benefits was refined to include only the best indicator or indicators for each benefit category, a method for developing the membership functions for the fuzzy expert system was also necessary. For the purposes of the SWBVA, it was determined that utilizing subjective evaluation and elicitation for development of the membership values was best. This method was chosen, because it involves the use of experts in the field who contribute their knowledge about what the membership values should be. These experts would, in fact, have similar backgrounds and experiences as the future users of the system and would provide the most realistic information on which to base the membership functions. Survey 2 consisted of two tables which listed all appropriate indicators from survey 1. Some indicators are viewed as beneficial through their decrease (i.e. a decrease in ‘lead time’ is considered a benefit), and some indicators are more beneficial when increased (i.e. an increase in ‘market share’ is considered a benefit). This is the reason for differentiating the tables between increases and decreases. The respondent was then asked to fill in the tables by entering what they considered to be a high, moderate, and low percent decrease or increase for each indicator. The expert’s opinions were collected in terms of a percentage for two reasons: (1) so that the survey results would be normalized among various companies

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

169

who have different product types and therefore differing production requirements and (2) because the membership values were modeled in ‘general’ terms of the values for these increases/decreases rather than a specific, calculated number. This makes the results more consistent among all potential users of the system and also reduces the amount of time and effort necessary for the expert to complete the survey. Response rates for surveys 1 and 2 were 11 and 45.45%, respectively. Out of 200 total survey 1 sent successfully, 22 were completed. Survey 2 was sent only to those who responded to survey 1, thus out of 22 successfully sent surveys, 10 were completed and returned. Results of the two surveys are outlined in the next section on fuzzy modeling procedures.

5. Results 5.1. Fuzzy modeling procedures Now that the data basis for the fuzzy modeling procedures have been established, an explanation of the steps taken to model the data into a fuzzy expert system can be provided. It was felt that fuzzy set theory was the best tool for accomplishing the objectives of this particular application because of the nature of the information involved with AMT investments and the use of decision maker perceptions that are necessary in determining the outcomes of the system. The ambiguity and vagueness involved in the types of system-wide benefits that are considered in the SWBVA make the nature of the information fuzzy. It is felt that the system wide benefits considered here are those which can often be intangible and therefore difficult to accurately put in dollar values. In addition, the need to provide some form of consistency between human perceptions in the evaluation process makes the use of a fuzzy expert system quite helpful. Lastly, the ability to easily use linguistic variables as inputs in a fuzzy system contribute to the SWBVA’s ease and comfort for both the user and interpreter of the system. Although a comprehensive overview of fuzzy set theory and fuzzy expert systems is not provided here, Cox (1994), Bezdek (1993), and McNeill and Thro (1994) provide excellent explanations of the principles and techniques involved. A general depiction of the fuzzy expert system component is shown in Fig. 4. Five main steps are involved in modeling a fuzzy expert system (McNeill and Thro, 1994). 1. Define the input variables for the system and their corresponding ranges of values. 2. Define the output variables for the system and their corresponding ranges of values.

Fig. 4. General depiction of a fuzzy expert system.

170

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

Table 2 Input and output variables as a result of survey 1 Sub-benefit category (output variable)

Best indicator(s) (input variables)

Increased product flexibility Increased process flexibility Increased demand flexibility Increased equipment flexibility Increased product quality

Shorter cycle times and set-ups Improved scheduling procedures Reduced lead times Increased machine efficiency Lower defect rates Reduced scrap and rework

Increased process quality Increased labor productivity Increased material productivity Increased capital productivity Increased overhead productivity

Improved control measures Decreased labor costs Decreased materials costs Increased planning accuracy Reduced inventories Increased process automation

Increased potential for product innovations Increased potential for process innovations Inner-departmental use of advanced technologies Fewer psychological barriers to automation Increased ability to achieve strategic goals Increased corporate strategy alignment Response to customers ahead of competitors Prevention of competitors’ gains in market share Quick response to customer demand

Reduced time to develop new products Number of processes that can be supported Increased communication within the system Increased learning about advanced technologies Number of achieved goals Increased progress on strategic goals Increased customer satisfaction Increased market share Reduced delivery time Reduced lead times

Improved customer image Increased market share for existing products Capture of new markets with future products Improved compatibility within the system Increased ability to integrate adjacent operations Improved management of operations Faster response to problems Enhanced employee buy-in to the system Improved quality of work life

Increased customer satisfaction Faster response to market demands Faster product introduction Standardization of product designs Shorter cycle times Increased communication within the system Availability of real-time information Increased employee morale/satisfaction Increased employee morale/satisfaction

3. Develop fuzzy membership functions for every input and output. 4. Develop a rule base based upon the potential outcomes of the system. 5. Determine how each action will be carried out by establishing the rule strengths and defuzzification. Steps 1 and 2 of this process were carried out through surveys 1 and 2. Each sub-benefit is considered as an individual output variable in the SWBVA. The narrowing of indicators for each system wide sub-benefit through survey 1 established the inputs for each sub-benefit. Table 2 outlines these results. The results of survey 2 then established the ranges of values to be considered for each input. These results are demonstrated in Table 3. One standard range was established for each of the output variables, which is based upon a percent increase rating from 0 to 100%.

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

171

172

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

Table 4 Sample rule matrix for increased process flexibility Increase in improvement in scheduling procedures None None

Low Low

Moderate Moderate

High High

Thus, the values for each of the input variables as entered into the system by the decision maker will eventually be translated into a crisp result for each output variable in terms of what percent increase they can expect from the acquisition of the particular AMT being evaluated. This crisp output value will then be compared to the benefit goals established by the decision maker in the calibration procedures of the tool to see if the desired amount of benefit has been obtained. Once the ranges for each input and output variable were established, it was possible to model these values into actual membership shapes. Each linguistic input and output name was developed into its own fuzzy set and corresponding membership function. Thus, each input and output variable has four membership functions in its universe, or domain of possible values: a set for ‘none’, ‘low’, ‘moderate’, and ‘high’. The ranges of values determined by the experts for each of these sets have been modeled into triangular membership shapes. Fig. 5 depicts some sample membership functions for the indicator category decrease in cycle time. Once the membership functions were created for the input and output variables, rule bases were assembled for translating the appropriate inputs into actual output values. Developing the rule bases involves first creating matrices for the inputs and then using these to establish the output actions that are most appropriate for a given situation. The matrix cells are then used as a basis for writing the actual rules that govern the system outcomes. In fact, each matrix cell represents one rule in the rule base. Because most of the output variables in the SWBVA have only one input variable, the establishment of the outcome actions was fairly straightforward. In these cases, a one-to-one relationship was used in determining the outcomes based upon the experts’ opinion that these input variables are the best indicators for the output variable in question. Two samples of this type of rule matrix for the output variables increased process flexibility and increased ability to integrate adjacent operations along with the rule base can be seen in Tables 4–6. For those output variables that have two input variables to consider, a similar criterion was utilized for establishing the outcome actions. Because actual inner-dependencies between the various input variables were not considered by the experts, the outcome actions for these situations were established as if the inputs were completely Table 5 Sample rule matrix for increased ability to integrate adjacent operations Decrease in cycle time None None

Low Low

Moderate Moderate

High High

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

173

174

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

Table 6 Rule base for increased ability to integrate adjacent operations 1

As you have no decrease in cycle time Then ability to integrate adjacent operations is none

2

As you have low decrease in cycle time Then ability to integrate adjacent operations is low

3

As you have moderate decrease in cycle time Then ability to integrate adjacent operations is moderate

4

As you have high decrease in cycle time Then ability to integrate adjacent operations is high

independent. Thus, the final outcome action listed for these situations will always consist of the highest possible value between the two input variables. The final step in the fuzzy modeling procedures, the process of defuzzification, involves utilizing the previously defined input and output variables, their associated membership functions, and rule bases to actually produce crisp output values based upon some type of user input. This detailed process consists of a several step process of mathematical manipulation of the defined components of the system. A detailed account of this process can be found in Yager and Filev (1994).

6. Case study and discussion: an example The detailed example presented here attempts to outline the process of SWBVA from start to finish while conveying as closely as possible the real experience that a decision maker would have in using the system. For this particular example, an Amada® laser system was chosen as the AMT to be evaluated. Amada America Inc. is a well-known manufacturer of laser systems as well as many other types of advanced manufacturing equipment. The particular laser system chosen for this example is an Amada 5 × 10 2000 W machine. Because the operating costs of a laser depend greatly on the type and thickness of the material processed, one main type of material had to be chosen for which to base the cost information. In this case, a material type of 0.125 in. stainless steel was chosen. The size of the laser system and the type of material to be processed were chosen, because they represented the “middle of the line” costs out of all the machine and material types outlined by Amada. The remainder of this section follows the steps of the SWBVA process for the Amada® laser system. First, the economic analysis procedures are explained and the IRR and associated x-gap results are presented. Then, the initial calibration procedures are completed, followed by the fuzzy input procedures based on the characteristics of an Amada laser system. Finally, the decision criteria for this particular example are analyzed and presented and a summary of the example are given.

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

175

176

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

Table 8 Economic analysis summary of the Amada 5 × 10 laser system Investment period

Net cash flows ($)

0 1 2 3 4 5 6 7 8

445000.00 5176.16 17176.16 59176.16 86176.16 108176.16 125176.16 142176.16 149176.16

Summary statistics IRR (%) 8.04

MARR (%) 15.00

x-Gap (%) 6.96

6.1. Economic analysis procedures The economic analysis procedures were performed independently based on the cost and benefit information provided by Amada. These cost and benefit categories are shown in Table 7. The SWBVA model was also used to check the possibility of quantifying some of those benefits for inclusion in the economic evaluation. Using information in Table 7, the IRR and the NPV of the net cash flows were calculated based upon a MARR of 15% and a planning period of 8 years. The net cash flows and the results of the economic analysis can be seen in Table 8. By examining this table, it is evident that the Amada laser system is not economically justified by the measurable costs and benefits and a positive x-gap value does exist. According to the SWBVA process, this indicates that further analysis is necessary in order to justify the purchase of the laser system. Some of the variables chosen for this example, such as the number of periods for which the investment is analyzed and the MARR, are completely arbitrary. In a real-world scenario, the values selected for such variables would be left entirely up to the individual company and the decision makers performing the analysis. 6.2. Calibration procedures Now that the economic analysis has been performed for the laser system and an x-gap value of 6.96% has been calculated, the next step in the SWBVA involves performing the calibration procedures to determine what benefits are useful for consideration and to determine goals for those benefits. At this point, the user of the system is presented with a list of the possible sub-benefit categories and the calibration questions. A depiction of this process with highlighted answers for the laser system example can be seen in Fig. 6. Based upon these answers, those benefit categories that are determined not useful for consideration in the adoption of an Amada laser system will be eliminated from consideration. In addition, the scales for the benefit goals will internally register to values between 0 and 100%. These values have been hypothetically estimated based on the location of the scales in Fig. 6 and are summarized in Table 9.

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

Fig. 6. Calibration procedures for the Amada laser system example.

177

178

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

Fig. 6. (Continued).

Table 9 Benefit goal translations for the Amada laser system example Output variable (sub-benefit category)

Benefit goal translation (% rating)

Increased product flexibility Increased process flexibility Increased equipment flexibility Increased product quality Increased process quality Increased capital productivity Increased potential for product innovations Increased potential for process innovations Increased ability to achieve strategic goals Response to customer demands ahead of competitors Prevention of competitors’ gains in market share Improved customer image Enhanced employee buy-in to the system Improved quality of work life

47 83 92 99 80 74 86 37 42 59 62 35 63 32

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

Fig. 7. Fuzzy input procedures for the Amada laser system example.

179

180

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

Table 10 Fuzzy user input translations for the Amada laser system example Input variable (indicator category)

User input translation (% increase or decrease)

Ability to improve scheduling procedures Defect rate scrap and rework rate Ability to improve control measures Cycle time Setup time Machine efficiency Time to develop new products Customer satisfaction Employee morale/satisfaction Planning accuracy Number of process innovations that can be supported Number of achieved strategic goals Market share

60 60 28 70 35 70 53 7 13 36 38 62 14 15

6.3. Fuzzy input procedures After the completion of the calibration procedures, the fuzzy input procedures are triggered. The user input questions and answers for this example are depicted in Fig. 7. From these inputs, translations would then be made internally within the system that correspond to the domain values of the membership functions as defined in Section 4. For the purposes of this example, estimates were made as to what these translations would be according to the placement of the scales. Using this method obviously does not generate a completely accurate translation, but the implementation of the SWBVA into a computer-programmed tool would eliminate the inaccuracies associated with performing these processes by hand. The estimated translation input values can be seen in Table 10. Table 11 Crisp output results for the Amada laser system example Output variable (sub-benefit category)

Crisp output value (% rating)

Increased product flexibility Increased process flexibility Increased equipment flexibility Increased product quality Increased process quality Increased capital productivity Increased potential for product innovations Increased potential for process innovations Increased ability to achieve strategic goals Response to customer demands ahead of competitors Prevention of competitors’ gains in market share Improved customer image Enhanced employee buy-in to the system Improved quality of work life

89 90 90 89 90 86 35 90 45 48 79 48 89 89

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

181

At this point, the generated user inputs would then be manipulated through the defuzzification processes referred to in Section 4. The details of these procedures will not be regenerated here, but the crisp output results have been summarized in Table 11. Detailed defuzzification procedure is provided in Mulvaney (1998). 6.4. Decision criteria A summary report detailing the results of the analysis for the Amada laser system can be seen in Fig. 8. From this, the decision-maker can examine what system wide benefits associated with the investment decision have fulfilled their benefit goals and what benefits have not met the justification requirements for the x-gap. Also outlined here are the overall

Fig. 8. Summary report for the Amada laser system example.

182

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

average values for each benefit category and one final system wide benefits value. From the summary report, it is apparent that only the goals for four sub-benefit categories have not been met: increased equipment flexibility, increased product quality, increased potential for product innovations, and response to customer demands ahead of competitors. At this point, the user of the system can weigh all of the information provided and then decide whether a readjustment of the benefit goals is possible based upon the value of the x-gap. If the user feels a readjustment is possible, he/she may do so and repeat the analysis. 6.5. Example summary The purpose of this case study example was to better exemplify the steps that a decision maker would take to evaluate their advanced technology investment decisions using SWBVA. Notice that the SWBVA does not attempt to explicitly tell users of the system what decision to make, but formally guides them through the process and allows them to think more intensely about the factors involved with such a decision. The SWBVA uses a base of expert knowledge along with the user’s own perceptions to provide some suggestions about the value of the system wide benefits they can expect to obtain from a particular AMT.

7. Implications and directions for future research More and more, small manufacturers are coming to the recognition that investment in advanced technologies is necessary for survival. Growing competition and increasing customer requirements and demands are the main factors contributing to this need. However, in order to justify AMTs, these companies must often consider system wide benefits in their analyses. This is due to many problems involved with traditional economic analysis of such equipment, including high hurdle rates and a tendency to compare investments with the status quo. The process proposed here, SWBVA, provides a formalized model of system wide benefits for decision makers to consider. It uses an expert base of knowledge to determine a more realistic value for the system wide benefits associated with a particular investment decision, while also allowing the decision maker to set their own goals for how much of those benefits they feel are necessary to justify the technology. The SWBVA makes the process of using the more intangible, system wide benefits in an investment evaluation more formalized, organized, and scientific. The use of a fuzzy expert system also allows for consistency among different decision makers. Several opportunities for further R&D have been identified in this area. First, it is recommended that the SWBVA process proposed in this research be implemented into a software or Internet-based tool. This will make the steps involved in the fuzzy evaluation procedures of the process much more accurate, and will allow the process to be easily accessed and manipulated by those that can most benefit from the use of such a tool. In addition, it is proposed that further efforts be made to add to the data basis of the system through additional surveys of the experts in this field. Although some data have been gathered to define the basic processes of the SWBVA, the gathering of more data to add to the fuzzy expert system will further enrich the validity of the tool. Another aspect to be considered in the collection of further data is the examination of the dependencies

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

183

between the input variables for those benefit categories that have more than one input. These dependencies were not taken into consideration in the development of the rule outcomes even though some interactions may in fact exist. The collection of further data could also include a surveying of the experts to determine what the proper outcomes should be when dependencies are taken into consideration. An additional area of research lies in the expansion of the SWBVA to also include the more intangible or system wide costs associated with AMTs. Christenson (1997) provides a good basis in this area by detailing the instances when newer technologies contribute to the failure of companies because of the unforeseen costs associated with them. The inclusion of a system wide costs model in conjunction with the system wide benefits model established here would make the evaluation process for advanced technologies even more complete.

Acknowledgements This study was supported by grants from National Science Foundation and Advanced Manufacturing Institute at Kansas State University. References Albayrakoglu, M., 1996. Justification of new manufacturing technology: a strategic approach using the analytic hierarchy process. Production and Inventory Management Journal 37 (1), 71–76. Azzone, G., Bertele, U., 1991. Planning and controlling investments in computer-based automation. Cost Management 28–36. Badiru, A.B., Foote, B.L., Chetupuzha, J., 1991. A multiattribute spreadsheet model for manufacturing technology justification. Computers and Industrial Engineering 21 (1–4), 29–33. Bennett, E.D., Hendricks, J.A., 1987. Justifying the acquisition of automated equipment. Management Accounting 69 (1), 39–46. Bezdek, J., 1993. Fuzzy models — what are they, and why? IEEE Transactions on Fuzzy Systems 1 (1), 1–5. Canada, J.R., Sullivan, W.G., 1990. Persistent pitfalls and applicable approaches for justification of advanced manufacturing technologies. Engineering Costs and Production Economics 18, 247–253. Christenson, C.M., 1997. The Innovator’s Dilemma When New Technologies Cause Great Firms to Fail. Harvard Business School Press, Boston. Cox, E., 1994. The Fuzzy Systems Handbook: A Practitioners Guide to Building and Maintaining Fuzzy Systems. AP Professional, Cambridge. Datta, V., Sambasivarao, K.V., Kodali, R., Deshmukh, S.G., 1992. Multi-attribute decision model using the analytic hierarchy process for justification of manufacturing systems. International Journal of Production Economics 28 (2), 227–234. Dhavale, D., 1995. Justifying manufacturing cells. Manufacturing Engineering 115 (6), 31–37. Downing, T., 1989. Eight new ways to evaluate automation. Mechanical Engineering 111 (7), 82–86. Finnie, J., 1988. The role of financial appraisal in decisions to acquire advanced manufacturing. Accounting and Business Research 18 (70), 133–139. Kakati, M., Dhar, U.R., 1991. Investment justification in flexible manufacturing systems. Engineering Costs and Production Economics 21 (3), 203–209. Kaplan, R., 1986. Must CIM be justified by faith alone? Harvard Business Review 64 (2), 87–95. Kassicieh, S.K., Ravinder, H.V., Yourstone, S.A., 1993. Proposed design of a DSS for the justification of advanced manufacturing technologies. IEEE Transactions on Engineering Management 404, 398–402. MacStravic, E.L., Boucher, T.O., 1992. A Software for capital investment analysis in manufacturing. Computers and Industrial Engineering 23 (1–4), 417–421.

184

S.M. Ordoobadi, N.J. Mulvaney / J. Eng. Technol. Manage. 18 (2001) 157–184

McNeill, F.M., Thro, E., 1994. Fuzzy Logic: A Practical Approach. AP Professional, New York. Meredith, J.R., Suresh, N.C., 1986. Justification techniques for advanced manufacturing technologies. International Journal of Production Research 24 (5), 1043–1057. Meredith, J.R., Hill, M.M., 1987. Justifying new manufacturing systems: a managerial approach. Sloan Management Review 28 (4), 49–61. Mohanty, R.P., Venkataraman, S., 1993. Use of the analytic hierarchy process for selecting automated manufacturing systems. International Journal of Operations and Production Management 13 (8), 45–57. Mulvaney, N., 1998. Development of a justification tool for advanced manufacturing technologies: system wide benefits value analysis. Master’s thesis, Kansas State University. Noaker, P.M., 1994. Can you justify change? Manufacturing Engineering 113 (6), 30–35. Noble, J.L., 1990. A new approach for justifying computer-integrated manufacturing. Cost Management 3 (4), 14–19. O’Brien, C., Smith, S.J.E., 1993. Design of the decision process for strategic investment in advanced manufacturing systems. International Journal of Production Economics 30/31, 309–322. Oeltjenbruns, H., Kolarik, W.J., Schnadt-Kirschner, R., 1995. Strategic planning in manufacturing systems — AHP application to an equipment replacement decision. International Journal of Production Economics 38, 189–197. Pant, S., Ruff, L., 1995. Issues in economic justification for flexible manufacturing systems and some guidelines for managers. Information Resources Management Journal 8 (1), 26–34. Parsaei, H.R., Wilhelm, M.R., 1989. A justification methodology for automated manufacturing technologies. Computers and Industrial Engineering 16 (3), 363–373. Parsaei, H.R., Karwowski, W., Wilhelm, M.R., Walsh, A.J., 1988. A methodology for economic justification of flexible manufacturing systems. Computers and Industrial Engineering 15 (1–4), 117–122. Presley, A., Barnett, W., Liles, D., 1995. A business case tool for the strategic justification of enterprise technologies. Computers and Industrial Engineering 29 (1–4), 421–425. Putrus, R., 1991. Accounting for intangibles in integrated manufacturing. Financial and Accounting Systems 30–35. Small, M., Chen, I., 1995. Investment justification of advanced manufacturing technology: an empirical analysis. Journal of Engineering and Technology Management 12, 27–55. Son, Y.K., 1993. Simulation-based manufacturing accounting for modern management. Journal of Manufacturing Systems 12 (5), 417–427. Son, Y.K., Park, C.S., 1987. Economic measure of productivity, quality, and flexibility in advanced manufacturing systems. Journal of Manufacturing Systems 6 (3), 193–207. Soni, R.G., Parsaei, H.R., Liles, D.H., 1990. A methodology for evaluating computer integrated manufacturing technologies. Computers and Industrial Engineering 19 (1–4), 210–214. Sullivan, W.G., 1986. Models IE’s can use to include strategic, non-monetary factors in automation decisions. Industrial Engineering 18 (3), 42–50. Swamidass, P.M., 1987. Planning for manufacturing technology. Long Range Planning 20 (5), 125–133. Swann, K., O’Keefe, W.D., 1990. Advanced manufacturing technology: investment decision process. Part 2. Management Decision 28 (3), 27–34. Troxler, J.W., Blank, L., 1989. A comprehensive methodology for manufacturing system evaluation and comparison. Journal of Manufacturing Systems 8 (3), 175–183. Wallsten, S.J., 1998. Rethinking the small business innovation research program. In: Branscomb, Keller (Eds.), Investing in Innovation. MIT Press, Cambridge, MA, pp. 194–220. Weber, S.F., 1993. A modified analytic hierarchy process for automated manufacturing decisions. Interfaces 23 (4), 75–84. Wilkes, F.M., Samuels, J.M., 1991. Financial appraisal to support technological investment. Long Range Planning 24 (6), 60–66. Xue, Y., 1999. Similarity-based reasoning using proverbs in managing technological innovation for small manufacturers. Master’s thesis, Kansas State University.