Sage Journals: Discover world-class research

Abstract

This article presents a closed-loop Kaizen system with discrete event simulation in the stage of design and reengineering of mass-customization production system. This approach contains four steps: current process analysis, production line design, discrete event simulation to solve the dynamic event questions, and implements and evaluation. The current process analysis and the production line design refer to the value stream mapping and the productivity analysis based on Kaizen principle, respectively. Production line designs are the modeling basis of discrete event simulation, and discrete event simulation feeds back results to the decision-makers to reduce waste and cost and improve the quality and productivity by reforming the production line. Therefore, by connecting the closed-loop Kaizen with discrete event simulation, the production line can be continuously and automatically improved. To demonstrate the potentials of the closed-loop Kaizen system, a case study was applied based on the original production layout and economic benefits were achieved. Moreover, this article indicates a direction for the full automation of process optimization.

Keywords

Closed-loop Kaizen system value stream mapping productivity analysis process optimization discrete event simulation

Introduction

With the increasing challenges from market competitions and environment regulations, manufacturers are placing more emphasis on the application of robot cells and enterprise information systems to improve the effectivity of production system and reduce the production cost.^1–3 At the same time, many manufacturers have adopted some advanced tools, methods, and techniques in their production activities. The tools and techniques containing value stream mapping (VSM), Lean–Kaizen, and the standardization of work have been used in production line^4–6 to process improvement. Continuous improvement is evolutionary improvement in quality management and manufacturing performance of the organizations for modern manufacturers.^7,8 Kaizen is a popular tool applied to eliminate wastes and minimize cost at all levels of production system,⁹ and it mainly aims to solve the real problems. The problems usually can be solved accurately because the results can be seen through the validation against a real scenario.¹⁰

The performance was optimized by the Bernoulli serial in the rework production activities under the ideal situation of machine.¹¹ In fact, discrete event simulation (DES) is an effective method of solving the stochastic problem and decreasing the fluctuation caused by uncertainties from the previous cells to make the production line more stable.^12–14 On the aspect of sustainability and forecast ability, DES can support decision making to determine the best scenario in a production line with stochastic variables and this procedure has been used for many years.¹⁵ In addition, a lot of good results in real production line have been reported by researchers in the energy optimization,¹⁶ product life cycle management,¹⁷ batch size optimization, and buffers optimization¹⁸ through the DES model. Due to the gap between the virtual simulation and the real production line, the optimization results are very different for different carriers. DES needs uncertainties, but how to avoid uncertainties and decrease the fluctuation caused by uncertainties from the previous cells to make the production line more stable becomes a new problem. Therefore, combining the merits of the continuous improvement methods and simulated the uncertainty event methods worth a further study.

This article presented process improvement by addressing the accurate optimization and continuous optimization problem. The main aim of this article is to develop a closed-loop Kaizen system by considering the effect of the process cell’s optimization and fluctuation of the cycle time in the production line. The proposed system uses VSM to detect problems that does not comply with Kaizen principle and DES to evaluate the performance of solutions. To assist the DES for modeling, productivity analysis is used and sensors are also used in this system to promote the accurate and continuous optimization by providing data for VSM and validate the execution effect of solutions. Due to the complexity of the manufacturing environment and the limit of robotic techniques, this article only focused on the production cell that can be reformed by using robots and the closed-loop Kaizen system is just robotization which is not fully intelligent.

This article is organized as follows. Section “Literature review” reviews related work. In section “Methodology,” the methodology of building a closed-loop Kaizen system is displayed. In section “Case study,” a case study was carried out to deeply optimize the process cell through the closed-loop Kaizen system with DES. The discussion is in section “Discussion.” Conclusions are presented in section “Conclusion.”

Literature review

The cores of process optimization are the application of intelligent devices and robotic cells, the integrated application of artificial intelligence and some lean management, and the continuous improvement of intelligent production system through simulation method. The development of production systems was a complex task, and researchers usually studied a variety of scenarios by productivity analysis especially in robotic cells. Hall et al.¹⁹ studied the sequence of a robot that produced similar parts in a manufacturing robotic cell with several machines. Geismar et al.²⁰ studied the problem of scheduling operations in robotic cells with dual gripper. Drobouchevitch et al.²¹ tried to maximize the productivity by analyzing the sequence. The robotic welding systems were developed in the shipyards to shorten the welding time.²² Gultekin et al.²³ found that a robot-centered layout could significantly reduce the cycle time.

In the past researches, the continuous improvement and lean management were focused by some scholars and manufacturers. The Kaizen method was widely used in the field of increasing operational efficiency, reducing cost, and improving the product quality to reduce cycle time and increase profits.²⁴ Glover et al.²⁵ proposed a Lean–Kaizen method to identify and eliminate wastes in production lines and provide continuous improvements in the small enterprise. Suarez Barraza et al.²⁶ illustrated that three techniques namely 5S (Seiri, Seiton, Seiso, Seiketsu, and Shitsuke), Gemba-Kaizen workshops, and process mapping which were related to Kaizen had a direct good effect on the processes and management systems. By the validation of reality, Kaizen method tightly connects with real manufacturing and can achieve improved effect as good as expected.

In the era of Industry 4.0, Rodič²⁷ presented a new simulation modeling paradigm to supply the seamless assistance in the lifecycle of production system supported by the real-time operation data. In the discrete manufacturing system, the process monitoring was achieved by the radio frequency identification to capture the real-time information for developing the scheduling simulation model.²⁸ The new simulation optimization tool of DES was used in many scenarios, such as prediction and evaluation of intractable production questions for the decision-makers to reduce the economic cost.²⁹ Turner et al.³⁰ constructed a test platform using DES to make decisions for smart factories. Fengque et al.³¹ found an intelligent DES architecture to optimize the decision by sharing resource information between different processes. Hoad and Kunc³² built a hybrid system combining the dynamic simulation and DES to offer a better understanding of real-world problems. DES depending on the powerful ability of computers can make the optimization better and more sustainable before the optimization is carried out.

Based on the contributions of past research, this article proposed a closed-loop method combining Kaizen and DES in production line to satisfy the needs of smart factory including the accuracy and sustainability of the optimization. Through this study, an application was carried out and expected results were achieved, which played an exemplary role for the wide use of this method. The research efforts will also enrich the current knowledge on process optimization.

Methodology

This article proposes a framework of closed-loop Kaizen system (as seen in Figure 1) to automatically and continuously optimize the production line. The procedure is shown as follows:

Step 1: Current process analysis. Through the ECRS (Eliminate, Combine, Rearrange, Simplify), SVM, action analysis, and Kaizen, the problems happening in production lines can be monitored and analyzed such as invalid action and idle time, time and motion study, manual operational analysis, and identified problems in process. VSM helps to display the most wasteful and unreasonable processes or key constraint problems and gives suggestions such as reducing machines, replacing workers by robots, and adjusting work sequence.

Step 2: Production line design. Depending on the current process analysis, the segment of production line design makes possible solutions such as production layout design, autonomation load/unload parts, robot design and choose, robot fixture design, and equipment linkage, by considering parameters such as cycle time, number of unqualified parts and qualified parts, idle time of machines, and arriving time of parts.

Step 3: DES. The DES model of production line is shown in Figure 2, including parts arrive segment, arrange machine segment, processing segment, and finish segment. The main simulation is the arrival time of parts, idle time of workers, loading and unloading time, and idle time of processing equipment. In the traditional production line, the manual operation brings fluctuation of cycle time and the robotic operation can conquer fluctuation of cycle time. However, in the smart production line, the robots dynamically respond to the time variation caused by the parts arrive segment in the production line. For example, the robots speed up to make up the late and keep the stability of production line during the parts arrive late. The model of DES is different for humans (Figure 2(a)) and robots (Figure 2(b)).

The last step is to verify the design effect through implementation and to feedback results to decision-makers to provide reference for the future optimization. Based on the analysis of DES results, the process optimization will be carried out by decision-makers and the effect of the optimization will be validated by the closed-loop system. If the reform is not well executed or some new problems happen, the defect optimization can be found immediately by sensors and DES model. The system will inform decision-makers again to improve the defect until the expected results are achieved.

Figure 1.

Framework of the closed-loop Kaizen system.

Figure 2.

Different operation flowchart: (a) manual operation flowchart and (b) robotic operation flowchart.

Case study

This method was used in the production line of air conditional compressor parts (ACCP) in a Chinese company B. The company B was facing the growing problems of external competition and internal personnel cost. They needed an automated and smart production line instead of the traditional production line to further reduce costs and improve the stability of production time.

Current process analysis

The current VSM (Figure 3) shown the existing processes and guides to identify problems for improvement. The upstream flow was for customers and the product was ordered monthly. The downstream flow was for supplier and the raw material was ordered monthly. The top flow was given at the top which represented information flow that moves from right to left between customer and supplier. The bottom flow was given at the bottom which represents the material moving from left to right in the ACCP production line. The production planning and control (PPC) department plans and monitors the quality status of the product in real time by Internet of things. The eight processes included the coarse grinding (P1), broaching (P2), middle washing (P3), surface grinding (P4), radius fine grinding (P5), hole polishing (P6), surface polishing (P7), and final washing (P8). The cycle time (CT), number of equipment (NE), number of workers (NW), unqualified rate every 10,000 (UR), and simultaneous processing quantity (SPQ) were recorded.

CT: the cycle time of a machine;

NE and NW: the number of equipment and worker respectively in a process cell;

UR: the unqualified rate of processed work pieces in every 10,000;

SPQ represents the number of processing work pieces that every worker has to manipulate simultaneously including loading, unloading, and carrying.

Figure 3.

Current value stream mapping of ACCP.

According to the VSM, it is obvious that P2 occupies the most people but has the highest unqualified rate, which is very unreasonable and needs to be improved. It is necessary and worthwhile to analyze the broaching (P2). The ECRS method is carried out to deeply analyze P2 and make preparation for the production line design and optimization from four aspects:

Elimination: eliminate invalid actions and idle time.

Combination: combine the operations to eliminate unnecessary time and action.

Rearrangement: rearrange the job order and resources to find the optimal sequence.

Simplification: simplify the operation by clamping three work pieces at a time.

Production line design

Through the current situation analysis, the production line was redesigned and two robots were planned to replace the three people, one of which operated two broaching machines and the other one of which operated another broaching machine (Figure 4). The coarse grinding cell was manual and thus the arriving time of work pieces for the net cell was uncertain. The three broaching machines did the same thing and work at full speed. Before executing the plan, it was necessary to simulate the production of humans and robots to assist decision making by illustrating advantages and disadvantages.

Figure 4.

Lean automatic production line layout of ACCP.

There were two types of robots to choose: (1) one with a six-claw manipulator and (2) the other one with a three-claw manipulator. The three types of work sequences were analyzed including manual work sequence. According to the theory of productivity analysis,³³ the best work sequence was shown in Figure 5. The manual work sequence was the same as the robotic sequence for one broaching machine (as seen in Figure 5(a)). The main difference between six-claw robot and three-claw robot was also that the unloaded and load time are different.

M, $M_{1}$ , and $M_{2}$ represent broaching machines;

$M_{0}$ , $M_{3}$ , and $M_{4}$ represent oil shedding machines which may be added to the process cell to reduce oil waste.

In the manual cell or one-machine cell of ACCP, the sequence (as seen in Figure 5(a)) starts with the movement of the robot at the input. The state of the system was expressed using ${e_{1}, e_{2}, e_{3}}$ , where $e_{1}, e_{2} ϵ {Ø, Ω}$ , and $e_{3} \in {I, O, M^{-}, M_{0}^{-}, M^{+}, M_{0}^{+}}$ . The state transitions in a cycle shown in Figure 5(a) could be expressed as follows

\begin{matrix} S_{1} = (Ω, ϕ, I) \to (Ω, ϕ, M^{-}) \to (Ω, ϕ, M^{+}) \\ \to (Ω, Ω, M_{0}^{-}) \to (Ω, Ω, M_{0}^{+}) \\ \to (Ω, Ω, O) \to (Ω, ϕ, I) \end{matrix}

S₁ starts with the movement of the robot at the input. Before the sequence starts, there should be work pieces being processed in M. Three work pieces were picked up and carried to M, where three old work pieces were unloaded and the three new work pieces were loaded. The robot moved to $M_{0}$ , where another three old work pieces had been processed last time. Work pieces in $M_{0}$ were unloaded and work pieces from M were loaded. The work pieces from $M_{0}$ were moved to the output where they were dropped. The robot moved back to the initial point and the cycle repeats.

Figure 5.

Work sequence of one and two broaching machines: (a) work sequence of one broaching machine and (b) work sequence of two broaching machines.

For production cell of ACCP with two broaching machines, the sequence (as seen in Figure 5(b)) starts with the movement of the robot at the input. The state of the system was expressed using ${e_{1}, e_{2}, e_{3}, e_{4}, e_{5}}$ , where $e_{1}, e_{2}, e_{3}, e_{4} \in {Ø, Ω}$ and $e_{5} \in {I, O, M_{1}^{-}, M_{2}^{-}, M_{1}^{+}, M_{2}^{+}}$ . By these notations, the state transitions in a cycle shown in Figure 5(b) could be expressed

\begin{matrix} S_{2} = (Ω, Ω, φ, φ, I) \to (Ω, Ω, φ, φ, M_{1}^{+}) \\ \to (Ω, Ω, φ, φ, M_{1}^{-}) \to (Ω, Ω, Ω, φ, M_{3}^{+}) \\ \to (Ω, Ω, Ω, φ, M_{3}^{-}) \to (Ω, Ω, φ, φ, O) \\ \to (Ω, Ω, φ, φ, I) \to (Ω, Ω, φ, φ, M_{2}^{+}) \\ \to (Ω, Ω, φ, φ, M_{2}^{-}) \to (Ω, Ω, φ, Ω, M_{4}^{+}) \\ \to (Ω, Ω, φ, Ω, M_{4}^{-}) \\ \to (Ω, Ω, φ, φ, O) \to (Ω, Ω, φ, φ, I) \end{matrix}

S ₂ starts with the movement of the robot at the input. Before the sequence starts, there should be work pieces being processed in $M_{1}$ and $M_{2}$ . Three work pieces were picked up and carried to $M_{1}$ , where three old work pieces were unloaded and the three new work pieces were loaded. The robot moved to $M_{3}$ , where another three old work pieces have been processed last time. Work pieces in $M_{3}$ are unloaded and work pieces from M₁ are loaded. The work pieces from $M_{3}$ were moved to the output where they were dropped. The robot moves back to the input. Three work pieces are picked up and carried to $M_{2}$ , where three old work pieces were unloaded and the three new work pieces were loaded. The robot moves to $M_{4}$ , where another three old work pieces had been processed last time. Work pieces in $M_{4}$ were unloaded and work pieces from $M_{2}$ are loaded. The work pieces from $M_{4}$ are moved to the output where they were dropped. The robot moved back to the initial point and the cycle repeats.

Based on the activity and behavior analysis, the sequence time of manual work and robotic work was measured and listed in Tables 1 and 2. It was defined that when the production line did not contain oil shedding machine, $tc$ equaled to 0. According to the measurement of this production line, $tg$ and $td$ varied from 6 to 7 s.

Table 1.

Robotic work sequence time.

Process	Idle	Get partsand moveto machine	Upload and download in machine	Upload and downloadin oil	Move to the output and drop the parts		Get parts and move to machine	Upload and download in machine	Uploadanddownloadin oil	Move to theoutput and drop theparts

Colors: busy status; $ta, tb, tc$ : time variables; oil: oil shedding machine; machine: broaching machine.

Table 2.

Manual work sequence time.

Process	Idle	Get parts and move to machine	Upload and download in machine	Upload and download in oil	Move to the output and drop the parts	Idle

Colors: busy status; $ta, tg, tb, tc, and td$ : time variables; oil: oil shedding machine; machine: broaching machine.

The time allocation for robotic production line scenarios including oil shedding machine is shown in Table 1. The first scenario was in the six-claw robot operation scenario: $tc$ equaled to 4 s, $tb$ equaled to 6 s, and it could be controlled to change from 5 to 7 s. The second scenario was in the three-claw robot scenario: $tc$ equaled to 5 s, $tb$ equaled to 7 s, and it could be controlled to change from 6 to 8 s.

The time allocation for the manual production line scenario including oil shedding machine was shown in Table 2: $tb$ varies from 8 to 10 s and $tc$ varies from 5 to 6 s.

DES for production scenario

In this case study, there were three production scenarios of ACCP, that is, manual operation, three-claw robots’ operation, and six-claw robots’ operation. Every time three parts arrived on each line and the interval time was about 17 s which was uncertain. DES method programmed by using C++ on computer was used to simulate the production of the three scenarios. The flowchart of DES program is shown as Figure 6. The uncertainties including arriving time and manual operation time were random and they could be produced by random function: srand ((unsigned) time (NULL)).

Figure 6.

Flowchart of DES program.

Implementation and evaluation

Through the implementation of DES, we got the average waiting time (Figure 7), utilization rate of machines (Figure 8) and workers (Figure 9), and fluctuation of cycle time (Figure 10) in the different scenario.

Figure 7.

Average waiting time in different scenarios.

Figure 8.

Utilization rate of machines in different scenarios.

Figure 9.

Utilization rate of workers in different scenarios.

Figure 10.

Fluctuation of cycle time in different scenarios: (a) previous cell scenario, (b) human cell scenario, (c) three-claw robot cell scenario, and (d) six-claw robot cell scenario.

The average waiting time of manual operation was the minimum and the three-claw robot operation was the maximum. The values increased when the oil shedding machines were added. The humans had the biggest utilization rate of machines and the six-claw robots had the smallest, which also illustrated that the machines can have more rest in the six-claw robot cell at the same productivity level. The utilization rate of broaching machines was the same for the same type of cell, no matter if there were shedding machines added. For the utilization rate of workers, the three-claw robots were the busiest and the humans were the idlest. When the oil shedding machines were added, all types of workers became busier but the value of three-claw robot 1 ascended to 94.1%, which was not good for the life.

The fluctuation of cycle time in the previous cell varies from 50 to 52 s which would influence the cycle time of the next process cell. The fluctuation of the manual increased from 47.3 to 55 s. Under the robotic operation cell scenario, the fluctuation was controlled, and the cycle time was limited between 50 and 51 s through the closed-loop Kaizen system.

After the application of oil shedding machines, the oil waste was eliminated. To test its effect, 10 work pieces were continuously weighed before and after the application of oil shedding machines. The comparison of deoiling effect is shown in Table 3.

Table 3.

Comparison of deoiling effect.

Sample	1	2	3	4	5	6	7	8	9	10
Undeoiling	1150 g	1158 g	1153 g	1149 g	1155 g	1152 g	1147 g	1156 g	1146 g	1152 g
Deoiling	1144 g	1154 g	1148 g	1143 g	1150 g	1149 g	1141 g	1152 g	1140 g	1149 g

Paired t-test was used to determine the effect of deoiling. The data were of normal distribution and the variance was equal. The results are shown in Figure 11. The effect of throwing oil was remarkable, and the average weight was 4.8 g lighter after throwing oil.

Figure 11.

Comparison chart of t-test.

The recycling of the oil was between 3 and 5 g for every product of different types and the average was 4g. The oil price was about 11.69 RMB/kg and the production output was 5.75 million pieces every year.

Total weight of the oil could be saved per year

4 g \times 5.75 million pieces = 23, 000 kg

Total cost of the oil could be saved per year

23, 000 kg \times 11.69 RMB / kg = 268, 870 RMB

The oil waste was successfully eliminated without decrement of productivity. At the same time, the unqualified rate also decreased from 1 to 0.1 every 10,000 through this method.

Discussion

From the results, it was obvious that the closed-loop Kaizen system could overcome the fluctuation brought by previous cell and made the production line more stable. Represented by manual, the non-dynamic system accumulated the fluctuation and made the production line more unstable. The average waiting time for manual was low and the utilization rate of broaching machines was good, but the utilization rate of manpower was too low. At the same time, the cost of manpower was 276,000 RMB/year but the cost of robots was 112,000 RMB/year. Three workers were replaced by two robots; the labor cost could be reduced by 59%.

The ability of controlling fluctuation of cycle time was the same for three-claw robots and six-claw robots. The average waiting time in six-claw robot cell was smaller than in three-claw robot cell. Although the utilization rate of machines in six-claw robot cell was 2% smaller than in three-claw robot cell, the utilization rate of robot in six-claw robot cell was 4% smaller than in three-claw robot cell, which said that the six-claw robot had more time to deal with added machines. When the oil shedding machines were added to the cell, the busy rate of six-claw robot only increased from 70.6% to 86.3%, but the three-claw robot’s busy rate was increased to 94.1%, which was too high. Therefore, the six-claw robot should be chosen.

When the oil shedding machines were added to the process cell, the average waiting time of work pieces would increase a little but the change could be neglected as to the cycle time. The utilization rate of broaching machines would not change and the utilization rate of robot would still be in reasonable range. Moreover, the cycle time of the process cell would not be affected, if the oil shedding machines were added. Therefore, the oil shedding machines were safe to use.

Based on the analysis of DES results, the process optimization was carried out by the decision-makers and the effect of the optimization was validated by the closed-loop system. Through twice-cycle running of the system, the process optimization of the production line achieved the expected results including the stability of the cycle time.

Conclusion

This article proposed a closed-loop Kaizen system with DES capable of continuously improving the production line and automatically controlling the fluctuation of the production line caused by uncertainties. The contributions of this article were summarized as follows:

A framework integrating the merits of Kaizen method and DES method to continuously optimize the production line was proposed.

A closed-loop Kaizen system which had the capability of automatically controlling the fluctuation of the production line caused by uncertainties was built.

A half-automatic system which simulated the results of reform and validates the results of optimization was displayed. A direction for the full automation of process optimization was indicated.

The feasibility of the proposed approach was validated by a case study. Through the application of the closed-loop system, a successful process optimization was displayed. The replacement of humans, the adoption of six-claw robots, and the addition of oil shedding machines were all indicted and validated by the system, and the fluctuation of cycle time was also controlled in real time.

Although the study demonstrated potentials of the closed-loop Kaizen system with DES, there are still more research that need to be done. The method is only based on the actual situation of this study. Some more complex applications and rules need to be further explored. Future research will focus on the automation of the closed-loop Kaizen system.

Footnotes

Appendix 1 Acknowledgements

The authors would like to thank Shanghai Institute of Producer Service Development (SIPSD), Shanghai Research Center for industrial Informatics (SRCI2).

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the National Natural Science Foundation of China (Grant No. 71632008).

ORCID iDs

Yuanju Qu

Siqi Qiu

Xianyu Zhang

References

Mourtzis

Vlachou

Milas

, et al. A cloud-based, knowledge-enriched framework for increasing machining efficiency based on machine tool monitoring. Proc IMechE, Part B: J Engineering Manufacture 2019; 233: 278–292.

Ming

, et al. An integrated framework of enterprise information systems in smart manufacturing system via business process reengineering. Proc IMechE, Part B: J Engineering Manufacture 2018; 233: 2210–2224.

Chen

. An Internet of Things based framework to enhance just-in-time manufacturing. Proc IMechE, Part B: J Engineering Manufacture 2017; 232: 2353–2363.

Arya

Choudhary

. Assessing the application of Kaizen principles in Indian small-scale industry. Int J Lean Six Sigma 2015; 6: 369–396.

Prashar

. Redesigning an assembly line through Lean-Kaizen: an Indian case. TQM J 2014; 26: 475–498.

Kumar

Dhingra

Singh

. Process improvement through Lean-Kaizen using value stream map: a case study in India. Int J Adv Manuf Tech 2018; 96: 2687–2698.

Bhuiyan

Baghel

. An overview of continuous improvement: from the past to the present. Manage Dec 2005; 43: 761–771.

Singh

. Continuous improvement approach: state-of-art review and future implications. Int J Lean Six Sigma 2012; 3: 88–111.

Vamsi Krishna Jasti

Kodali

. A literature review of empirical research methodology in lean manufacturing. Int J Oper Prod Manage 2014; 34: 1080–1122.

10.

Karim

Arif-Uz-Zaman

. A methodology for effective implementation of lean strategies and its performance evaluation in manufacturing organizations. Bus Process Manage J 2013; 19: 169–196.

11.

Zhou

Lin

. A novel analysis model for optimization performance of Bernoulli serial production system considering rework processes. Proc IMechE, Part B: J Engineering Manufacture 2019; 234: 295–309.

12.

Kang

Bhatti

. Continuous process improvement implementation framework using multi-objective genetic algorithms and discrete event simulation. Bus Process Manage J 2019; 25: 1020–1039.

13.

Lins

Droguett

. Redundancy allocation problems considering systems with imperfect repairs using multi-objective genetic algorithms and discrete event simulation. Simul Model Prac Theor 2011; 19: 362–381.

14.

Baril

Gascon

Miller

, et al. Use of a discrete-event simulation in a Kaizen event: a case study in healthcare. Eur J Oper Res 2016; 249: 327–339.

15.

Trigueiro de Sousa Junior

Barra Montevechi

de Carvalho Miranda

, et al. Discrete simulation-based optimization methods for industrial engineering problems: a systematic literature review. Comput Ind Eng 2019; 128: 526–540.

16.

Ghani

Monfared

Harrison

. Energy optimisation in manufacturing systems using virtual engineering-driven discrete event simulation. Proc IMechE, Part B: J Engineering Manufacture 2012; 226: 1914–1929.

17.

Morshedzadeh

Oscarsson

, et al. Multi-level management of discrete event simulation models in a product lifecycle management framework. Proc Manuf 2018; 25: 74–81.

18.

Barrera-Diaz

Oscarsson

Lidberg

, et al. Discrete event simulation output data-handling system in an automotive manufacturing plant. Proc Manuf 2018; 25: 23–30.

19.

Hall

Kamoun

Sriskandarajah

. Scheduling in robotic cells: complexity and steady state analysis. Eur J Oper Res 1998; 109: 43–65.

20.

Geismar

Dawande

Sriskandarajah

. Robotic cells with parallel machines: throughput maximization in constant travel-time cells. J Schedul 2004; 7: 375–395.

21.

Drobouchevitch

Sethi

Sriskandarajah

. Scheduling dual gripper robotic cell: One-unit cycles. Eur J Oper Res 2006; 171: 598–631.

22.

Yan

Fang

Ong

, et al. Optimal pass planning for robotic welding of large-dimension joints with nonuniform grooves. Proc IMechE, Part B: J Engineering Manufacture 2018; 232: 2386–2397.

23.

Gultekin

Akturk

Karasan

. Scheduling in robotic cells: process flexibility and cell layout. Int J Prod Res 2008; 46: 2105–2121.

24.

Dora

Van Goubergen

Kumar

, et al. Application of lean practices in small and medium-sized food enterprises. Br Food J 2014; 116: 125–141.

25.

Glover

Farris

Van Aken

, et al. Critical success factors for the sustainability of Kaizen event human resource outcomes: an empirical study. Int J Prod Econ 2011; 132: 197–213.

26.

Suarez Barraza

Smith

Mi Dahlgaard-Park

. Lean-Kaizen public service: an empirical approach in Spanish local governments. TQM J 2009; 21: 143–167.

27.

Rodič

. Industry 4.0 and the new simulation modelling paradigm. Organizacija 2017; 50: 193–207.

28.

Yuan

Guo

Wei

, et al. Radio frequency identification-enabled monitoring and evaluating in the discrete manufacturing process. Proc IMechE, Part B: J Engineering Manufacture 2016; 231: 2184–2196.

29.

Huang

Hsieh

, et al. Simulation optimization in the era of Industrial 4.0 and the industrial Internet. J Simul 2017; 10: 310–320.

30.

Turner

Hutabarat

Oyekan

, et al. Discrete event simulation and virtual reality use in industry: new opportunities and future trends. IEEE T Hum-Mach Syst 2016; 46: 882–894.

31.

Fengque

Yifei

Fei

, et al. Research on design of the smart factory for forging enterprise in the industry 4.0 environment. Mechanics 2017; 23: 146–152.

32.

Hoad

Kunc

. Teaching system dynamics and discrete event simulation together: a case study. J Oper Res Soc 2018; 69: 517–527.

33.

Shafiei-Monfared

Salehi-Gilani

Jenab

. Productivity analysis in a robotic cell. Int J Prod Res 2009; 47: 6651–6662.

Process optimization through closed-loop Kaizen with discrete event simulation: A case study in China

Abstract

Keywords

Introduction

Literature review

Methodology

Case study

Current process analysis

Production line design

DES for production scenario

Implementation and evaluation

Discussion

Conclusion

Footnotes

Appendix 1

Acknowledgements

Declaration of conflicting interests

Funding

ORCID iDs

References