Introduction
Since the beginning of 2020, the Covid-19 epidemic has had a major impact on most industries. However, demand was increasing for the medical garment industry, opening up great opportunities for businesses in this industry. Instead of being able to take advantage of the opportunity to increase customers, MZ company (a company in medical garment industry in Vietnam) (MZC), lost a significant number of orders.
The 5-Why method was applied to find the initial causes of the lost-order problem. The initial causes were high production costs and low service levels. The company faced cost problems. An increased cost in resources leads to an increased cost in production that affects their market share. Manufacturing resources needed to be used effectively to cut costs.
The company’s current service level, ability to meet orders on time, was quite low. The average order delay rate in the first 10 months of 2020 is about 64.8%. A fishbone diagram combined with the brainstorming method was used to analyze the main causes of late orders, as shown in Figure 1.
Figure 1. Fishbone diagram for the possible causes of late orders.
Using the failure mode and effects analysis (FMEA) method to identify and prioritize the causes of late orders in the production of the company. The result with risk priority number (RPN) percentage P(%) and RPN cumulative percentage CP(%) was shown in Table 1.
Table 1. Analyzing and prioritizing possible causes.
The Pareto chart for the main causes of late orders was shown in Figure 2.
Figure 2. The Pareto chart for the causes of late orders.
The six main causes of late delivery with an RPN cumulative percentage of more than 80% were inappropriate production plans, improper production schedules, unspecific forecasting models, incorrect material requirement planning (MRP), wrong order promising, and inconsistent bills of materials. The six causes with corresponding solutions were shown in Table 2.
Table 2. Causes analysis and solutions.
The solutions led to a manufacturing resource planning (MRP)II, which consists of four functional modules as shown in Table 2. The MRPII system was designed to reduce late orders and then increase the customer service level. The system also reduced resource costs, thereby helping to solve the problem of lost orders for the company. The MRPII system was constructed for the SS product family of the company, including six models: SS1, SS2, SS3, SS4, SS5, and SS6.
Literature review
According to (1), the MRPII is a system for effective planning of all resources of a manufacturing company. It is made up of a variety of interlinked functions such as Demand Management (DM), Production Planning (PP), Master Production Schedule (MPS), MRP, Capacity Requirement Planning (CRP), and Vendor Requirement Planning (VRP). The MRPII is an approach to managerial planning, execution, and control of productive activity. It integrates, in a feedback manner, the forecasting of demand, PP, scheduling and control, and purchase planning and control.
Amad-Uddin et al. designed a bespoke MRPII system for an small and medium enterprise (SME) company with the critical modules DM, MPS, MRP, and CRP (2). Phong et al. developed an MRPII system in the electric generator manufacturing industry with modules of DM, PP, MPS, and MRP (3).
Purna Chandra Padhan used Seasonal Autoregressive Integrated Moving Average (SARIMA), a time series forecasting model, in forecasting cement productions. The forecasting performance of various competing models was evaluated through forecast accuracy criteria mean absolute percentage error (MAPE) (4). Emiro Antonio Campo et al. proposed and implemented an aggregated PP model to provide optimal strategies in the medium term for a textile company. Linear programming was used to minimize total costs associated with labor and inventory levels. The model took into account characteristics associated with fabric contraction, wastes in the process, the efficiency of new employees, and training requirements (5).
Jia-Nian Zheng and Chen-Fu Chien developed a system for excelling enterprise resources for light-emitting diode (LED) manufacturing that optimizes chip procurement and PP by linear programming (6). P.J. Weeda constructed a stochastic model for the reduction of the initial forecast in the master schedule of an MRP system during the progress of time by the acceptance of customer orders (7).
Raqeyah Jawad Najy found out that an overstated MPS caused raw materials and work in process (WIP) inventories to increase and led to missed due dates. He used rough cut capacity planning (RCCP) to validate the MPS with respect to available capacity (AC) (8). Juin-Han Chen and Chin-Tai Chen proposed a two-phase order promising process in which available to promise (ATP) was first reserved in phase I to meet the most important demand. Customer orders were promised in Phase II according to the time-phase manufacturing resource supply calendar and restricted via the ATP in Phase I (9).
Teeradej Wuttipornpun et al. developed a practical finite capacity MRP system based on the needs of an automotive-part manufacturing company in Thailand. The proposed system offered a good trade-off between conflicting performance measures and resulted in the best weighted average performance measure when compared with conventional forward and forward-backward finite capacity scheduling systems (10).
Research methodology
The research methodology for developing the MRPII system includes the following steps.
Step 1: Develop DM module
Step 2: Develop PP module
Step 3: Develop MPS module
Step 4: Develop MRP module
Develop demand management (DM) module
This module has to achieve the objective of accurately forecasting customer demand. The DM module has two inputs: the long-term demand, or historical demand, and short-term demand, or customer order. Historical demand data in the product family will be used to make a demand plan for the PP module by suitable forecasting models. Forecasting models are also used to forecast product mix percentages for each product in the product family. Customer orders will go through the DM module to the MPS.
Forecasting is done according to the following steps:
1. Determine forecast parameters
2. Collect data
3. Eliminate outliers in the data
4. Select appropriate forecasting models
5. Forecast according to the best model.
Forecast parameters include forecast objects and forecast times. The forecast object can be a product family or individual product. The forecast time is 1 year with a cycle of 1 month. Past data is collected to forecast the forecast object. The data is processed, and outliers are eliminated to ensure the reliability of forecast data. Data patterns are analyzed to select appropriate forecast models. The best forecast model is selected according to forecast errors.
Develop production plan (PP) module
The PP module receives the demand plan from the DM module and constructs the production plan of the product family to minimize the cost of using the resources and constraints on resource capacity (1). Production plans are constructed by using operations research models in (11).
The PP steps are as follows.
1. Identify resource alternatives.
2. Construct the model
3. Collect data, estimate model parameters
4. Determine the production plan.
Resource alternatives used to meet demand in the case include regular-hours production, overtime production, inventory, and back-order. The model is constructed by defining model indexes and parameters, thereby determining the objective and constraints.
The index and parameters used in this case are defined as shown in Table 3.
Table 3. The model index and parameters.
The decision variables used in this case are defined in Table 4.
Table 4. The model variables.
To minimize the total resource cost, the objective function is defined as follows.
Min C
The constraints are constructed as follows with t = 1÷12.
1. Demand constraint: X(t) + Y(t) + I(t−1) – I(t) + B(t) – B(t−1) = D(t)
2. Regular hour production capacity constraint: X(t) ≤ RC(t)
3. Overtime production capacity constraint: Y(t) ≤ OC(t)
4. Inventory constraint: 0 ≤ I(t) ≤ MI
5. Backorder constraint: 0 ≤ B(t) ≤ MB*D(t)
6. Inventory & Back-order relation: I(t)*B(t) = 0
7. Initial & ending inventory: I(1) = II, I(12) = EI
8. Initial & ending backorder: B(1) = IB, B(12) = EB
9. Non-negativity restrictions: X(t), Y(t), I(t), B(t) ≥ 0
After the model is built, data is collected to estimate model parameters. The model is solved to determine production plans for the product family.
Develop master production schedule (MPS) module
The MPS module makes a production schedule for each product to meet the production plan of the product family (1) in three steps.
1. Step 1: Design the MPS
2. Step 2: Make the MPS
3. Step 3: Control the MPS
Design the MPS
Designing MPS determines scheduling objects, scheduling time, and scheduling parameters. The scheduling objects are products within the product family. Scheduling time is monthly with daily cycles. The scheduling parameters include the number of production days per month (DPM), consumption periods for adjusting MPS in step 2, and time fences for controlling MPS in step 3, as in Table 5.
Table 5. The MPS parameters.
Make the MPS
Making MPS includes the following steps.
1. Draft an initial production schedule from forecast demand.
2. Adjust the production schedule according to actual customer orders.
3. Make the production schedule feasible.
Draft an initial production schedule from the forecast demand. The MPS receives a demand forecast (DF) for each product from the PP module. The DF for each product is derived from the production plan for the product family and the product mix percentage of each product, forecasted by the DM module. The monthly forecasted demand for each product will be divided equally among each day to create an initial production schedule.
Adjust production schedule according to actual customer orders. The MPS also receives actual orders from the DM module. The forecast consumption model (1) will be used to revise the production schedule as customer orders consume the forecast from the production plan. When there are orders that exceed forecast demand, the excess quantity will consume the unconsumed forecast (UF) quantities in the previous cycles (backward consumption) and the following cycles (forward consumption). After consumption is completed, the revised production schedule (RPS) will be defined according to the customer order (CO) and the UF, as follows.
Make the production schedule feasible. Making the production schedule feasible includes the following steps.
1. Determine the critical station and its profiles.
2. Determine the rated capacity (RC) of the station.
3. Determine the RC of the station.
4. Check the feasibility of the production schedule.
5. Adjust when the schedule is not feasible.
The critical station is often the bottleneck station in the production process with the lowest cycle time (CT). The station profiles include parameters of the stations as shown in the following Table 6.
Table 6. Critical station profile.
The rated capacity (RC) of the stations is determined as follows.
For the sake of simplicity, with the assumption that the CTs are the same for every product produced in the station, the required capacity is defined according to total demand (TD) per date, and the station CT, as follows.
The production schedule is considered feasible when RC is lower than AC at every scheduling period. If RC exceeds AC at some scheduling periods, loads in these periods will be reallocated to other periods to make the schedule feasible.
Control the MPS
The production schedule must also be controlled by using models of projected available balance (PAB), and ATP, to keep the MPS stable under demand fluctuations (1).
Projected available balance (PAB). Projected available balance (PAB) is used to evaluate the performance of the production schedule and whether scheduled production balances with future demand. With t as the time index of the MPS, the PAB is defined according to the time fences (DTF) and PTF, the beginning on-hand (BOH), the MPS, the CO, and the forecast (F) as follows:
1. 1 ≤ t ≤ DTF:PAB(t) = PAB(t − 1) + MPS(t) − CO(t),PAB(0) = OHB.
2. DTF < t ≤ PTF:PAB(t) = PAB(t − 1) + MPS(t) − Max[CO(t), F(t)].
Available to promise (ATP). Available to promise (ATP) is used to promise the customer orders at specific periods in the future. At first, discrete ATP without looking ahead DATP-WOL is defined according to the BOH, the MPS, and the CO before the next MPS. For periods that MPS = 0:
For periods that MPS ≠ 0:
1. t = 1 : DATP − WOL = OHB + MPS − ΣCO
2. t > 1 : DATP − WOL = MPS − ΣCO
Then discrete ATP with look-ahead DATP-WL is defined according to discrete available to promises - without look-ahead (DATP-WOL). In periods that DATP-WOL is less than 0, the ATP will be set to 0, and the over-promised amount will be compensated by taking the ATP of the nearest previous periods that have positive ATP. Finally, the ATP is defined by cumulating the discrete ATP with a look ahead.
Develop material requirement plan (MRP) module
The MRP module determines the requirement of materials based on the structure of the products and the inventory status of each material to meet the MPS (1). The steps for planning material requirements are as follows:
1. Determine the production schedule.
2. Determine the product structure.
3. Determine the inventory status of materials.
4. Determine the MRP.
The production schedule is identified from the MPS module. Product structure is often presented through a Bill of Materials (BOM). The inventory status of materials is determined by the inventory status record (ISR), which includes parameters as shown in Table 7.
Table 7. Inventory Status Record (ISR) parameters.
Material requirements are planned through variables as shown in Table 8.
Table 8. MRP variables.
Material requirements are planned over time through the following steps (1).
1. Determine the gross requirement (GR).
2. Determine the net requirement (NR).
3. Determine the planned order receipt (PORc).
4. Determine the projected on-hand (POH).
5. Determine the planned order release (PORl).
The GR is determined by the production schedule for the final product and by the PORl of the corresponding parents of the material. The NR, PORc, the POH, and the PORl are determined as follows.
The research methodology will be applied to the case in the following sections.
Demand management (DM)
The DM module used for demand forecasting was included.
1. DF for product family SS.
2. DF for individual products in the product family.
Demand forecast for product family
To forecast the demand for the product family of SS in the next 12 months of the next year, 2021, SS demand for the last 3 years, 2018, 2019, and 2020, was collected as shown in Table 9.
Table 9. Demand of SS for the last 3 years.
Processing the data showed that the data set has no outliers. After analyzing both the trend and seasonal behavior of the data set, the proposed models were auto-regressive integrated moving average (ARIMA), Decomposition, and Winter. Minitab software was used to forecast the three methods. Mean absolute deviation (MAD) and tracking signal (TS) were used as the forecast errors to choose the most suitable model. The forecast errors of the proposed models were shown in Table 10.
Table 10. Forecast errors of the proposed models.
The Winters method was selected. The forecast result was shown in Table 11.
Table 11. Demand D of SS for the next year, 2021.
Demand forecast of each product
To determine the DF of each of the six products SS1, … SS6, the demand percentages (DP) of all products were collected in the previous year. The DP of SS1 was forecasted, and the remaining products were done similarly. The DP of SS1 in the last 2 years is shown in Table 12.
Table 12. Demand percentage (DP) of SS1 in 2019, 2020.
Since the data were collected random and not trending, forecasting methods of Naive Method (NM), Moving Average (MA), and Exponential Smoothing (ES) were proposed. The forecast errors of the proposed models were shown in Table 13.
Table 13. Forecast errors of the proposed models.
The method of MA was selected. The forecast result was shown in Table 14.
Table 14. The DP for each product in 1/2021.
The demand for each product in 1/2021, D(1), is forecasted by the corresponding DP(1) and PP(1) as shown in Table 15.
Table 15. The demand D(1) for each product in 1/2021.
Production planning (PP)
The PP model, as in section “Develop production plan module PP,” was constructed as follows.
Min C
St.
The monthly demand of the SS product family D(t), t = 1–12, was forecasted according to Table 11. Data were collected, and model parameters were estimated. Regular and overtime production capacity in units of products per month in the next year are estimated as shown in Table 16.
Table 16. Production capacity in 2021.
Parameters of costs, inventory, and backorder were estimated as shown in Table 17.
Table 17. Production costs, inventory, and backorder parameters.
Applying the Open-Solver add-in in Excel, the production plan in units of products for the next year, 2021, was found in Table 18.
Table 18. Production plan in 2021.
With the above production plan of product family SS, the total cost of resources was minimal as follows.
Master production schedule (MPS)
Design the MPS
The production schedule for all models of the product family was made daily for a month. The SS1 would be selected to make its production schedule, and the production schedules of other products were made similarly. The month chosen to be scheduled is January 2021. The MPS parameters were defined in Table 19.
Table 19. The MPS parameters.
January has 31 days, of which 25 are production days, as in Table 20.
Table 20. 31 days in January.
The 25 production days of January are shown in Table 21.
Table 21. The 25 production days in January.
Make the MPS
Draft the initial production schedule
The DF of SS1 in January was 396, as shown in Table 15. This DF was equally distributed for the production days of the month to make the initial production schedule, as in Table 22.
Table 22. The initial production schedule of SS1.
Revise the production schedule according to actual customer orders
Along with the initial MPS, or the DF, were the CO, as shown in line CO in Table 23. The forecast consumption model was used to consume DF by CO to make the revised MPS. The model applied backward forecast consumption first, then forward forecast consumption later, and the number of days for the consumption was 6 days, as in Table 19. The result with the consumed forecast (CF), the UF, and the revised MPS was shown in Table 23.
Table 23. The MPS of SS1 after consumption.
The revised MPS for model SS1 daily in January 2021 was shown in Table 24.
Table 24. The revised MPS of SS1 in January 2021.
Similarly, the revised MPSs for the remaining models were made as shown in Table 25, with the TD of all models per day.
Table 25. The revised MPS of all models in the SS family in January 2021.
Make the MPS feasible
The MPS was checked for feasibility by applying the resource profiles method in (1). The SS production process consists of nine stations: M0, M1, …, and M8. According to station CT, the critical station in the process was M5. The parameters of the M5 station were all estimated and shown in the station profiles shown in Table 26.
Table 26. M5 station profile.
The AC of M5 in minutes per day was calculated as follows.
The AC of M5 in the three dates of maintenance was calculated as follows.
The AC of M5 in January was shown in the AC lines of the load report in Table 27. For the sake of simplicity, with the assumption that the CTs are the same for every product, the RC in minutes per date was defined according to TD per date and the CT of M5, as follows:
Table 27. M5 load report in January 2021.
The required capacity (RC) in January to meet the TD in Table 25 was shown in the RC lines in Table 27. The station load report of the M5 with AC, RC, and overload (OL) was shown in Table 27, where the OL was defined as follows.
The M5 station was overloaded on days 6, 13, 22, and 27 and under load on the other days, as shown in Figure 3.
Figure 3. The load diagram of M5.
The load on the OL days would be reallocated on the days of underload to make the MPS feasible. The MPS after load reallocation was shown in Table 28.
Table 28. The revised MPS in January after load reallocation.
The M5 load report after load reallocation was shown in Table 29 and Figure 4. The OLs of all planning periods were now all negative; the MPS was now feasible.
Table 29. The M5 load report after load reallocation.
Figure 4. The load diagram of M5 after load reallocation.
Control the MPS
To control the MPS, PAB, and ATP were calculated from the MPS, DF, and CO; the initial OH of 2; the demand time fence (DTF) of 5 days; and the planning time fence (PTF) of 12 days. The PAB and ATP for model SS1 were calculated and shown in Table 30. The MPS of other models could be controlled similarly.
Table 30. The PAB & available to promise (ATP) for model SS1.
Material requirement planning (MRP)
Material requirement planning (MRP) determines the materials plan based on the BOM of the products, and the ISR of each material to meet the MPS (1). SS1 was selected as the representative to do the MRP; other products were done similarly.
The MPS of SS1 was shown in Table 28. The BOM of SS1 was shown in Table 31. The materials with units of pieces (P), or rolls (R), included two types of purchasing (P), and manufacturing (M).
Table 31. BOM of SS1.
The ISR of manufacturing materials was shown in Table 32.
Table 32. The ISR of manufacturing materials.
The ISRs of purchasing materials were shown in Table 33.
Table 33. The ISR of purchasing materials.
From the BOM and ISR, use the MRP add-ins in Excel to plan the material requirements that meet the MPS. The plan for manufacturing materials was shown in Table 34.
Table 34. The plan order release for manufacturing materials.
The plan for purchasing materials was shown in Table 35.
Table 35. The plan order release for purchasing materials.
The plan of manufacturing materials would be the input for CRP. The plan of purchasing materials would be the input for VRP. CRP and VRP were beyond the scope of this study.
Conclusion
The article evaluated the current state of the resource planning system of the company and developed an MRPII system for the company under study with functional modules including DM, PP, MPS, and MRP.
Initially, forecasting models in the DM module were used to forecast the demand of the product family. The PP module developed a production plan that met the demand for the product family and the resource constraints. The MPS module developed production schedules for each product in the product family that matched the production plan. Finally, the MRP module developed material plans that met the production schedules.
The research integrated these functions to reduce production costs and improve customer service levels, then reduce late orders. The article has the following advantages.
1. Applied forecasting techniques to accurately forecast customer demand.
2. Applied operations research to minimize the cost of using resources when developing a production plan.
3. Used the forecast consumption method to adjust the production schedule.
4. Checked the feasibility and controlled the MPS through PAB and ATP.
5. Used the ATP model to promise orders efficiently to increase service level.
6. Established MRP that met the MPS.
However, the research still had some disadvantages. It was only studied and implemented on a product family and just stopped at the planning level. Operations research did not apply to capacity planning. The parameters of the MPS module were chosen by experience. Furthermore, it had not been performed in a practical environment to evaluate the effectiveness of the system. The system would be complete with CRP and VRP. These restrictions lead to further research to improve the system and move towards the ERP system.
Funding
The authors declare that no financial support was received for the research, authorship, and/or publication of this article.
Conflict of interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
References
1. Nhu PN. Manufacturing Resource Planning MRP II. Vietnam National University Publisher (2020). Available online at: https://www.smashwords.com/books/view/1044194
2. Amad-Uddin Khan MK, Noor S. Design & implementation of a bespoke MRPII system for a small and medium enterprise (SME) manufacturing company. J Qual Technol Manag. (2011) VII(I):73–90.
3. Nhu PN, Vo THT, Duy MP. Developing Manufacturing Resource Planning MRP II System. A Case Study in HT Company. (2018). Available online at: www.isem.edu.vn
4. Padhan PC. Use of univariate time series models for forecasting cement productions in India. Int Res J Fin Econ. (2012) 83. Available online at: http://www.internationalresearchjournaloffinanceandeconomics.com
5. Campo EA, Cano JA, Gómez-Montoya RA. Linear programming for aggregate production planning in a textile company. Fib Text East Eur. (2018) 26(5(131)):13–9. doi: 10.5604/01.3001.0012.2525
6. Zheng J-N, Chien C-F. Master production schedule and system for excelling enterprise resources (SEER) in the LED industry. IEEE International Conference on Automation Science and Engineering (CASE). (2013). Available online at: https://ieeexplore.ieee.org/document/6654010
7. Weeda PJ. A stochastic model for forecast consumption in master scheduling. Int J Prod Econ. (1994) 35:401–4.
8. Najy RJ. Rough cut capacity planning-RCCP-case study. Adv Theoret Appl Mech. (2014) 7(2):53–66. doi: 10.12988/atam.2014.4612
9. Chen J-H, Chen C-T. Using mathematical programming on two-phase order promising process with optimized available-to-promise allocation planning. Int J Comput Int Manag. (2009) 17(3):25–40.
10. Wuttipornpun T, Yenradee P, Beullens P, van Oudheusden DL. Finite capacity material requirement planning system for a multi-stage automotive-part assembly factory. ScienceAsia. (2006) 32:307–17. doi: 10.2306/scienceasia1513-1874.2006.32.307
11. Nhu PN. Operations Research. Smashwords ebook (2020). Available online at: https://www.smashwords.com/books/view/1052246
© The Author(s). 2025 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.