Quantitative vs. Qualitative method
Qualitative and quantitative research methods have long been two main research methodologies among academia. Qualitative research is a method to explore and understand the meaning individuals or groups give to a social or human problem. The research includes the process of emerging questions and procedures, collection of data typically in the participant’s settings, inductive analysis of data moving from particulars to general themes and making interpretation of the data by the researcher (Creswell, 2009).
Quantitative research at the other hand, is a tool for testing objective theories through examining the relationship among variables. These variables, in turn, can be measured and turned into numbered data that can be analyzed using statistical procedures. The final report of this research method should have a structure consisting introduction, literature and theory, methods, results and discussion. Those involve in this type of research are interested in deductive analysis and testing of theories, evaluating alternative explanations and being able to generalize and replicate the findings (Creswell, 2009).
There are a set of differences between these two traditions. The most important difference between them is the way in which each tradition treats data (Brannen, 1992). In quantitative approach, the researcher tries to test a theory by specifying and narrowing down a hypotheses and by collecting data to support or refute the hypotheses (Creswell, 2009). In theory, if not in practice, the researcher defines and isolates variables and variable categories. The variables then, are linked together to frame hypotheses often before the data is collected, and are then tested upon the data (Brannen, 1992). The qualitative researcher at the other hand, begins with defining very general concepts which will change in their definitions as the research progresses. For the former, the variables are the tools and means of the analysis while for the latter, they are the product or outcome of the research (Brannen, 1992). As an example, in qualitative method, the researcher tries to establish the meaning of a phenomenon from the views of participants. This requires to identify a culture-sharing group and to study how it develops shared patterns of behavior over time (Creswell, 2009). The qualitative researcher is said to look through a wide lens, looking for patterns of inter-relationship between a set of concepts that are usually unspecified while the quantitative researcher looks through a narrow lens at a set of specified variables (Brannen, 1992).
The second important difference between the two methods is the way they collect data. In the qualitative tradition, the researcher must use himself as the instrument, attending to his own cultural assumptions as well as to the data. In order to gain insights to the participants’ social worlds the researcher is expected to be flexible and reflexive and yet manufacture some distance (Brannen, 1992). Qualitative approach includes three main kinds of data collection methods: in depth, open-ended interviews; direct observation; and written documents (Johansson, 1995).
In quantitative tradition, the instrument is a finely tuned tool which allows for much less flexibility, imaginative input and reflexivity, for example a questionnaire. By contrast, when the research issue is less clear and questions to participants may result in complex answers, qualitative methods like in-depth interviewing may be called for (Brannen, 1992). Compared to qualitative method, the main quantitative research techniques include the use of questionnaires, structured interviews, measurement, standardized tests, statistics and experiments (Johansson, 1995).
Qualitative approach studies selected issues in depth and detail. This is due to the ability to approach the fieldwork with openness and without being constrained by predetermined categories of analysis. On the other hand, quantitative methods require the usage of standardized methods so that the wide variety of perspectives and experiences of people can be fitted into a small number of predetermined response categories. The most advanced method in quantitative research is experiment where fieldwork is replaced by laboratory (Johansson, 1995).
In quantitative approach, the researcher often tries to minimize the effects of intervening factors on the research phenomenon. In qualitative approach, the researcher tries to find out and describe what the intervening factors are and how they influence the research phenomenon under study (Johansson, 1995).
In quantitative research, the researcher works with statistics and uses the average, the frequency, the causality and the prediction as a base for the report. In qualitative research, the researcher believes that if something has happened once, it can happen again even if you cannot calculate where and when (Johansson, 1995).
As a research strategy, case study has been used in many different situations to contribute to our understanding and knowledge about individual, group, organizational, social, political, and related phenomena. Case study is even used in economics, where the structure of a given industry or the economy of a given region or city is investigated by case study techniques. In all of these cases, the need for a case study arises out of the desire to understand complex phenomena. In brief, the case study allows the researcher to retain the holistic and meaningful characteristics of real-life events (Yin, 2003).
Case study is defined as:
“An empirical inquiry that investigates a contemporary phenomenon within its real-life context, especially when the boundaries between phenomenon and context are not clearly evident. The case study inquiry copes with the technically distinctive situation in which there will be many more variables of interest than data points, and as one result relies on multiple sources of evidence, with data needing to converge in a triangulating fashion, and as another result benefits from the prior development of theoretical propositions to guide data collection and analysis” (Yin, 2003, p.13).
In other words, you use case study method because you deliberately want to cover contextual conditions believing that they are highly important to your phenomenon of study. Second, because phenomenon of study and its context are not always distinguishable in real-life situations, a whole new set of technical characteristics like data collection and data analysis strategies is required (Yin, 2003).
A case study research can include both single-case and multiple-case studies. Although some fields have tried to distinguish sharply between these two approaches, they are in reality two variants of case study designs. A case study can also include or even be limited to quantitative evidence and as a related but important note, the case study strategy should not be confused with qualitative research (Yin, 2003).
In order to investigate the problem of “lot sizing” or “finding optimum batch sizes” in a high mix, low volume production environment a case company has been selected. The company is a manufacturer of different types of electronic products. To focus more on the problem, one of the main workstations of the company that produces different types of electronic boards using surface mount technology is chosen.
The case company chosen for this thesis report is Westermo Teleindustri AB, an electronics manufacturing company. Westermo was established in 1975. Its first data communication product was an RS-232 line driver, allowing data transmission over large distances using twisted pair cables. With its head office in southwest of Stockholm, it grew over the past three decades to establish subsidiaries in Sweden, UK, Germany, France, Singapore, North America, Taiwan with sales partners over 35 countries. In 1990s, Westermo created the world’s first industrial DIN rail mount telephone modem. Today it designs and manufactures robust data communication devices for harsh environments. With its strong commitment to develop its own industrial data communications solutions, last year it invested 13% of its turnover in R&D. Westermo’s ambition is to deliver a customer service level of 98% with return ratios below 0.25%. As a result, Westermo conducts business with a large number of system integrators around the world while having special partner programs with some of them (Westermo.com, 2014).
Amongst different products of the company are the printed circuit boards (PCBs). Today, up to 188 different boards are produced in the company. High variety of boards and low volumes classify the production as High mix, Low volume. The need for frequent long changeovers forces the production line to produce the boards in batches.These boards are used as a component in company’s other final products or they are delivered directly to the customers as finished products. The boards are produced in one of the company’s production lines using Surface Mount Technology (SMT). The SMT assembly involves three basic processes: screen printing of the solder paste on the bare boards, automatic placement of components on the boards using two placement machines in series (one for small components and the other for large components), and solder reflow oven. There are inspections after the solder printing, placement machines and reflow oven. The boards are produced in batches. Batch sizes are specified in an ERP system called IFS. Whenever customer demand cannot be met by finished boards in inventory, a production order of a specified quantity is sent to the workstation through IFS.
Research method, data collection and analysis
The nature of the batch sizing problem requires the description of the demand pattern, finding averages, dealing with large amount of numeric data and carrying on optimization procedures. Due to the nature of the research problem, it is necessary to continue with a quantitative approach.
At the beginning of the project, a thorough literature review was carried on on similar topics and articles in peer reviewed journals and in previous thesis works on relevant subjects. Articles from the university data base and the textbooks from the university library were the main sources of data. Afterwards, in order to make a better understanding of the problem at hand in detail levels, an investigation of the production process was performed through daily visits of the SMT line, making close observations, asking questions from operators and the production manager and searching relevant data through company’s data base.
The data required to solve the research problem was collected from company’s ERP system. This data includes information related to demand patterns for each board, prices, production quantities, cycle times, capacity and etc. The data from ERP system was in raw form and had to be processed before turning into meaningful information, therefore a great deal of time was spent on processing and manipulation of raw data using Excel. To continue, an optimization model was created which enabled this data to be used. The model was used not only for calculating the optimum batch sizes, but also to perform capacity analysis and investigating the effect of setup time reduction on both batch sizes and capacity.
In this chapter, the theoretic framework of this report is explained. Relevant theories are described and later used in the empirical part.
Description of an SMT line
Printed Circuit Boards (Figure 1) are the central part of an electronic product and are manufactured through automated assembly lines with one or several stations where necessary components are placed on the boards (Salonen, 2008). Surface-mount technology refers to assembling of the electronic components on boards by soldering them onto their surface where components are placed on one side or both sides of the board (Coombs, 2008). SMT technology can be traced back to 1960s when it was first used for assembling hybrid microcircuits (HMC). The surface-mount technology provides manufacturers with the ability to use smaller components and create greater densities on the boards (Coombs, 2008).
According to Coombs (2008), the main advantage obtained from surface-mount technology is lower manufacturing cost resulting from automated assembly processes. There are three basic assembly processes in an SMT line including (1) printing of solder paste on the boards, (2) placing the components on the boards, and (3) reflow the solder in a furnace (Coombs, 2008). Solder paste which is a combination of solder powder, thixotropic agents and flux is applied on the boards with great precision (thickness and area). One common method for applying solder paste is screen or stencil printing. In this method, the solder paste is applied on the boards through openings in the screen or stencil called apertures. The apertures are located on exact locations on the boards where solder paste is required (Coombs, 2008). Figure 2 is an example of a solder paste printing machine:
Pick and place machines can handle small or large electrical components and put them precisely where solder paste deposits are placed. The tacky nature of the flux in the solder paste keeps the components in place (Coombs, 2008). According to Salonen (2008), placing machines are classified as either gantry or turret style based on the design of the pick and place system. Gantry style machines have a number of nozzles on a movable placement head which can move between the feeder bank and component placement location on the board and can pick any component and place it on the board. Feeder banks and the boards are usually fixed and do not move during the placement process. In contrast, a rotary turret style machine has a fixed head and a movable feeder carrier that provides the next required component for the placement head and a movable table that holds the board in the exact placement position (Salonen, 2008). Figures 3 to 6 show two pick and place machines, a horizontal turret placing head and a feeder bank.
Table of contents :
1.2 Problem formulation
1.3 Aim and research questions
1.4 Project limitations
2. Research method
2.1 Quantitative vs. Qualitative method
2.2 Case study
2.3 Case company
2.4 Research method, data collection and analysis
3. Theoretic framework
3.1 Description of an SMT line
3.2 Economic Order Quantity
3.3 Discussion about setup cost
3.4 Single Minute Exchange of Die
3.4.1 Fundamentals of SMED
3.4.2 Discussion over SMED
3.5 Linear optimization, nonlinear optimization
3.6 Exact methods, Heuristics
3.6.1 Genetic Algorithm
4. Results (Empirics)
4.1 Calculating EOQs
4.2 Discussion over setup cost
4.3 Writing an optimization model
4.4 Solving the optimization model
4.5 Setup improvement
6. Conclusions and Recommendations
A. The GA code for MATLAB
B. Tables 12 to 14