9. CONCLUSIONS AND RECOMMENDATIONS
9.1.2 Current performance with uncontrolled stock points
The second sub-question is related to the performance of the current stock points. There is no data available about the inventory control policy of the current stock points, nor about the performance of the inventory model. Moreover, the yearly holding costs computed by Vanderlande are time independent. The performance in delivering order on time and complete to SCCE of VIM in week 2 to 37 of 2016 is 83% and of VIS in week 10 to 37 of 2016 is 65%, while the threshold for both factories is 95%.
The performance gap between the actual performance and the desired performance is therefore 12%
for VIM and 30% for VIS. These results are displayed in section 3.1.
In section 3.2, the causes are categorised. This brings to light that 35% of the delays are caused by overdue purchase parts. An analysis of the Delivery to request of the second tier suppliers, executed in section 3.4, showed that both suppliers that deliver to VIM, and suppliers that deliver to VIS are underperforming. The second tier suppliers of VIM have an average performance of 80%, and the suppliers of VIS have an average of 83.3%. The threshold for the Delivery to request performance of the second tier suppliers is 98%. The research addresses three factories and a subcontractor, but due to information unavailability only the performance of VIM and VIS and their second tier suppliers is evaluated.
To conclude, the second tier suppliers are underperforming in delivering material at or before the request date set by Vanderlande. Consequently, the material availability at the factories is low and this affects the performance of the factories.
54 9.1.3 Five local stock points and local decisions
The third sub-question investigates the impact of local decision making and local stock points. A conceptual design of a two-echelon distribution system is defined, and in section 5.3, demand data is analysed in order to estimate the demand parameter. The analysis in section 5.3 led to the conclusion that Vanderlande can determine PI demand for multiple items at the end of the conceptual design phase.
A PI demand analysis, executed in Chapter 6, shows to what extend Vanderlande can determine PI demand based on SPO Future demand. The PI analysis proved that nearly all PI demand is defined at the end of the conceptual design, and that Vanderlande should redesign its control structure, instead of stock raw and intermediate material. As a result, the third sub-question became irrelevant.
9.1.4 A single stock point at the second tier supplier and central decisions
The fourth sub-question also became irrelevant after the conclusion of the data analysis revealed that a redesign will provide a better solution for Vanderlande.
9.1.5 Five local stock points and central decisions
The fifth sub-question also became irrelevant after concluding that a redesign leads to a better solution.
9.1.6 Compare the performance of the proactively and non-proactively controlled stock points The goal of the sixth sub-question was to compare the inventory models, though due to the new insights, these models were not developed and the performance was not compared.
9.1.7 Effect of a redesign on the performance of the factories and suppliers
The objective of the redesign was to increase the performance of the factories by increasing the material availability. Therefore, the process structure, control structure, and decision structure are redesigned insection 7.4. Figure 24 illustrates the qualitative redesign. The main difference in the process structure, discussed in sub-section 7.4.1, involved transmitting information between sales engineering and SCCE.
This adjustment will save time and will result in better communication.
The redesigned control structure, elaborated on in sub-section 7.4.2, leads to more improvements. First, translating the SPO Future demand into PI Demand provides insight in the resources required for the production of sub-components. Moreover, the ordering and signalling of PIs by SCCE increases material availability as second tier suppliers have more time to produce material. In addition, central and weekly ordering of material results in larger batches and can lead to quantity discounts. Material coordination is increased by central stock points for the few items that still require inventory. Furthermore, the redesigned control structure can decrease product uncertainty, process uncertainty, complexity of structure of goods flow, and complexity of the multi-project character of the ETO situation.
The adjustments in the decision structure, discussed in sub-section 7.4.3, ask for closer cooperation between departments, and can decrease technical risks, increase workload control, and increase the flexibility for VIM.
55
Altogether, the redesigned process structure, control structure, and decision structure impact the performance of the factories, as they can decrease material unavailability and increase the flexibility of VIM, resulting in less capacity issues. The SPO dataset for purchase items consists of 76 items of which 37% are longer lead-time items that currently require expediting or arrive too late. This percentage is reduced to 3% by the redesign.
9.1.8 Other effects
There are some other effects that result from analysises performed during the research project.
Effect of lead-time uncertainty
In Chapter 8, the effect of lead-time uncertainty on the performance of inventory models is investigated.
A periodic review, single item, single location, inventory control policy with base-stock levels is used to simulate four types of models with either an adjusted or fixed replenishment level, and either deterministic or stochastic lead-time. The demand and ratios of two items is used to estimate a demand distribution used as input for the model. The first item is part of component (B) and has little demand with an average of 2.70 items per week, the second item is part of component (C) and has a larger demand, with an average of 64.43 items per week. The lead-time of the two stochastic models can be one week longer than the standard lead-time. Different probabilities are used as input to compute the effect of the variation. The computational results in sub-section 8.10.2 show that the models with an adjusted replenishment level perform slightly better than the models with a fixed replenishment level.
Moreover, the costs for the models with the adjusted level are a lot smaller.
Little lead-time uncertainty has a big impact on the service level of the models. For the first item, the impact was bigger on the model with an adjusted replenishment level and for the second item, the impact was similar between the items. The model with an adjusted replenishment level has lower costs, therefore resulting in better output for item 2. Furthermore, the effect of lead-time uncertainty is bigger on the models of the item with large weekly demand.
A small sensitivity analysis on the probability of an item in a component, conducted in sub-section 8.10.2, shows that little differentiation in the probability can have a big impact on the performance. Therefore, it is important to find the right probabilities.
9.2 Recommendations
This section elaborates on the recommendations to increase factory performance.
9.2.1 Redesign the control structure
In order to increase factory performance, it is recommended to redesign the process and control structure by increasing the interaction between departments, translating the SPO Future demand into PI demand, central ordering of PIs with a lead-time longer than four weeks, signalling of other items, and central inventory control, as it will increase material availability. Furthermore, it is recommended to redesign the order acceptance functions as they affect worload control and increase flexiblity. This solution will directly increase factory performance, as it tackles the two biggest causes (Appendix F).
56 9.2.2 Analyse other equipment
This master thesis project solely focused on SPO data as these projects are always produced by Vanderlande factories and due to the high future demand illustrated in the SPO Future demand. By analysing the list with all purchase items ordered in VIM and VIS, it is concluded that 15% of the items currently require extra effort to be delivered on time (Table 15). It is recommended to analyse other equipment similar to the analysis conducted in section 5.3 and Chapter 6, as this might lead to the conclusion the redesign is applicable to other equipment.
Table 15, Lead-time division of all purchase items
9.2.3 Reduce lead-time uncertainty
The results of the simulation of the inventory control policy showed that little uncertainty has a large impact on the performance of an inventory model. For the purchase items with a small demand, a fixed replenishment level is advised as it results in the best performance, whereas for purchase items with larger demand, an adjusted replenishment level results in similar performance as a fixed level, but has lower costs. It is suggested to decrease this uncertainty by close cooperation and information sharing with the second tier suppliers. Signaling demand is part of this closer cooperation and can reduce lead-time uncertainty. Moreover, lead-lead-time control can often be reduced by adding an additional crashing cost, or by long-term partnerships between suppliers and vendors (Ouyang et al., 2004).
9.3 Academic relevance
Little research is conducted on control policies in an ETO or capital goods industry. Bertrand and Muntslag (1993) analyzed production control in an ETO firm, but did not test their model in a real-life situation. This master thesis project contributes to literature by providing a case study on the model of Bertrand and Muntslag (1993). No evidence was found for literature that studies control policies for a supply chain with multiple factories. By this research, we have tried to fill this gap by using a business case study from a capital goods company.
Van Aken (2007) states that a typical research product of a design science is the technological rule: “A technological rule gives, for a specific solution concept, the objectives the application of the solution concept would serve, and for which setting it would be valid.” By this project, the conceptual control system of Bertrand and Muntslag (1993) is field-tested in its intended field of application.
9.4 Suggestions for further research
Multiple assumptions are used in order to scope the problem. For this reason, a first suggestion for further research is to relax these assumptions and analyse the effect of the relaxation on the results.
The validation of the redesign involved four aspects that ask for further research. First, the redesign is only feasible for SPO projects. The SPO is part of Parcel and Postal projects. There are multiple other
Items Ratio
Lead-time ≤ 4 weeks 5052 85%
Lead-time > 4 weeks 857 15%
57
projects that contain purchase items used for the SPO. Consequently, it is suggested to first analyse those projects and extend the redesign over those items, and lateron analyse other projects.
Secondly, in Chapter 7 and 8, it is assumed that the lead-time to America is equal to the lead-time in Europe. Relaxing this assumption will have a big impact, as the lead-time to America can be a couple of weeks longer due to extra transportation time. The new sub-order allocation rule defined in sub-section 7.4.3 will make it possible to estimate the amount of demand that is required in America, however Vanderlande first has to gain information about all different characteristics before it can implement the true situation in America. Shipping material asks for batch sizing and consolidation of items, two aspects that are not considered in this master thesis project. On top of that, multiple political issues arise when shipping material to America instead of buying it there. Therefore a suggestion would be to first analyze the control structure of the factory and the differences between Europe and America in a logistical but also political manner.
Another limitation of the research is the lack of quantitative support for the redesign. It is suggested to analyse the quantitative effect of a redesign of the processes, control structure, and decision structure.
The fourth aspect for further research involves the IT capabilities. It is validated whether it is possible to adjust the IT system, but the detailed adjustments are not clarified yet.
The validation of the inventory control policies involved some more aspects that ask for further research.
Firstly, an assumption that is made in the inventory control policy is that the lead-time for the purchase item is five weeks. However, the dataset of the purchase items shows that some items have a lead-time of 6, 7 or 8 weeks. The effect of these lead-times on the inventory model have to be investigated in a next research project.
The demand for the first item was rather small, therefore it is hard to draw conclusions from this output.
The second item had a larger demand, resulting in more thrustworthy output. However, only two items were used as input. The operational validity can be extended by adding more items to the model.
Moreover, only 25 weeks with data were used as input for the demand of component (B) and (C). This can be extended in further research, making the research more reliable.
The SPO Future demand currently contains information about all characteristics but three. If SCCE is able to add information about the three characteristics, it can decrease uncertainty in purchase items demand and thus inventory. This should be investigated in cooperation with sales engineering.
Moreover, the planning of the SPO Future demand is considered as deterministic in this research. In real-life the planning can deviate, as the customer might ask for acceleration of deceleration. The possibility for deviations to the control structure and inventory control policy will increase the fit to the structure of Vanderlande. Furthermore, a what-if scenario on multiple second tier suppliers for a specific purchase item can be executed.
58
59 BIBLIOGRAPHY
Adan, I., Van eenige, M., & Resing, J. (1995). Fitting discrete distributions on the first two moments.
Bertrand, J., & Muntslag, D. (1993). Production control in engineer-to-order firms. Eindhoven, The Netherlands: Elsevier.
Bertrand, J., Wortmann, J., Wijngaard, J., Suh, N., Jansen, M., Fransoo, J., . . . de Jonge, T. (2015).
Design of Operations Planning and Control Systems. Eindhoven: School of Industrial Engineering, Eindhoven University of Technology.
Bijvank, M. (2013). Periodic review inventory systems with a service level criterion. Journal of the Operational Research Society 65, 1853-1863.
Chopra, S., & Meindl, P. (2013). Supply Chain Management. Edinburgh : Pearson Education Limited.
Christopher, M. (2005). Creating value-adding networks. Logistics and Supply Chain Management.
Durlinger, P. (2013). Demand Management: Voorraadbeheer. Durlinger Consultancy.
Gosling, J., Towill, D. R., Naim, M. M., & Dainty , A. R. (2009). Engineer-to-order supply chain management: A literature review and research agenda. International journal of Production Economics, 741-754.
Hobday, M. (2000). The project-based organisation: an ideal form for managing complex products and systems? Science and Technology Policy Research, 871-893.
Iida, T. (2015). Benefits of leadtime information and of its combination with demand forecast information. International Journal of Production Economics.
Ishikawa, K. (1990). Introduction to quality control. Tokyo, Japan.
Law, A. M. (2007). Simulation modeling and Analysis (Fourth edition). Tucson, Arizona, USA: McGraw-Hill.
Mitroff, I. I. (1974). A methodology for strategic problem solving. Management Science.
Ouyang, L.-Y., Wu, K.-S., & Ho, C.-H. (2004). Integrated vendor-buyer cooperative models with stochastic demand in controllable lead time. International journal of production economics.
Radke, A. M., & Tseng, M. M. (2012). A risk management-based approach for inventory planning of engineering-to-order production. CIRP Annals - Manufacturing Technology, 387-390.
Sargent, R. G. (2013). An introduction to verification and validation of simulation models. Syracuse, NY, USA: Syracuse University.
Shrivastava, P. (1987). Rigor and practical usefulness of research in strategic management. Strategic Management Journal, 77-92.
Topan, E. (2014). Supply Chain Operations Planning, Distribution systems. Eindhoven.
60
Van Aken, J. E. (2007). Problem solving in Organizations, A methodological handbook for business students. Cambridge: Cambridge University Press.
Vanderlande. (2015). Annual report. Retrieved from https://www.vanderlande.com/about-vanderlande/annual-report
61 APPENDIX
Appendix A. List of Abbreviations Abbreviation Description
CDF Cumulative Distribution Function
EDC European Distribution Centre
EOQ Economic order quantity
ETO Engineer-to-order
ERP Enterprise Resource Planning
FPS Field Problem Solving
JDE J.D. Edwards World; Enterprise Resource Planning software
KPI Key performance indicator
LR Literature Review
MOQ Minimum Order Quantity
MRP Material Resource Planning
ON Inventory Order
PDF Probability Distribution Function
PI Purchase item
PIO Purchase item order
PO Production order
OF Factory Order (Veghel)
RP Research Proposal
SCE Supply Chain centre Europe
SPO Posisorter
TU/e Eindhoven University of Technology
V Vanderlande Industries
VIA Vandende’s factory in America
VIM Vanderlande’s factory in Veghel
VIS Vanderlande’s factory in Spain
VP Vice President
WK Work Kit
WO Work Order
62 Appendix B. Figures, Tables and Graphs
Figure 1, Redesign of the control structure ... V Figure 2, ETO trade-off by Radke and Tseng (2012) ...1 Figure 3, Product hierarchy ...5 Figure 4, Material and information flow ...6 Figure 5, Lead-time from factory order release until finish date ...7 Figure 6, Ishikawa diagram of factory performance ... 11 Figure 7, Layout of the Posisorter ... 12 Figure 8, Regulative cycle by Van Strien (1975) ... 16 Figure 9, Situation 1: Local decision making and local stock points ... 18 Figure 10, Situation 2: Central decision making and local stock points ... 18 Figure 11, Situation 3: Central decision making and central stock ... 19 Figure 12, Sub-components and types of Component (B) ... 21 Figure 13, Sub-components and types of component (C) ... 22 Figure 14, Sub-component groups of Component (F) ... 22 Figure 15, Adjusted Regulative Cycle ... 26 Figure 16, Current Control Structure ... 27 Figure 17, Procurement activity ... 30 Figure 18, Goods Flow Control and production phases... 30 Figure 19, Aspects of GFC at Vanderlande ... 34 Figure 20, Production control redesign ... 34 Figure 21, Customer order acceptance function ... 34 Figure 22, Sub-order assignment and PU outsourcing decision ... 35 Figure 23, WO release function ... 36 Figure 24, Redesign control structure ... 36 Figure 25, Relationship between different BOM types (Bertrand and Muntslag, 1993) ... 37 Figure 26, Relation between theoretical model and simulation ... 42 Figure 27, Simplified version of the model of development process (Sargent, 2013) ... 48 Figure 28, Variants of component (B) ... 69 Figure 29, Variants of component (C) ... 69 Figure 30, Variants of component (F) ... 70
Table 1, Performance measures second tier suppliers ... 13 Table 2, Comparison of suppliers VIM ... 14 Table 3, Key Performance Indicators set for Inventory Control ... 17 Table 4, Example demand analysis ... 20 Table 5, Components and their measurements ... 20 Table 6, Improvement possibilities ... 28 Table 7, Production Units at Vanderlande ... 32 Table 8, Holding cost percentage according to Durlinger (2013) ... 47 Table 9, Input values, item 1 and item 2 ... 49 Table 10, Replenishment levels, item 1 ... 50 Table 11, Results of model 1 and 3, item 1 ... 50 Table 12, Results of model 2 and 4, item 1 ... 50 Table 13, Sensitivity analysis, item 1 ... 52 Table 14, Sensitivity analysis, item 2 ... 52 Table 15, Lead-time division of all purchase items ... 56 Table 16, Ratios of variants of component (B) ... 71 Table 17, Ratios of variants of component (C) ... 72
63
Table 18, Ratios of variants of component (F) ... 73 Table 19, Overview of PI analysis ... 74 Table 20, Variants and items of component (B) ... 78 Table 21, Variants and items of component (C) ... 79 Table 22, Results fitting procedure (Adan et al., 1995) ... 80 Table 23, Operational validation item 1 ... 82 Table 24, Operational validation item 2 ... 82 Table 25, Replenishment levels item 2 ... 83 Table 26, Results of model 1 and 3, item 2 ... 83 Table 27, Results of model 2 and 4, item 2 ... 83
Graph 1, Performance and costs of an inventory model with lead-time uncertainty ... VII Graph 2, Weekly performance VIM ... 10 Graph 3, Weekly performance VIS ... 10 Graph 4, SPO Future demand in meters ... 12 Graph 5, Delivery to request, second tier suppliers VIM... 13 Graph 6, Delivery to request, second tier suppliers VIS ... 14 Graph 7, Average performance and cost comparison, item 1... 51 Graph 8, Delivery to request, second tier suppliers SPO VIM ... 68 Graph 9, Delivery to request, second tier suppliers SPO VIS ... 68 Graph 10, Historic demand of component (C)... 77 Graph 11, Warm-up period... 80 Graph 12, Validation of Binomial distribution ... 82 Graph 13, Average performance and cost comparison, item 2 ... 83
64 Appendix C. Assumptions
All assumptions discussed throughout the research are described in order of occurrence.
Chapter 3: Performance measures analysed in section 3.1 and 3.4 are reliable
The performance of the factories is controlled by SCCE and the performance is measured similar for all first tier suppliers. The performance is based on a date set by SCCE. Therefore, it might be considered
The performance of the factories is controlled by SCCE and the performance is measured similar for all first tier suppliers. The performance is based on a date set by SCCE. Therefore, it might be considered