Journal of Human Resource Management
Volume 3, Issue 2, April 2015, Pages: 6-16

Create Heijunka 5’s Matrix to Control the Dynamic Gemba as Lean Tool

Ahmed M. Abed

Industrial Engineering Department, Zagazig University, Zagazig, Egypt

Email address:

To cite this article:

Ahmed M. Abed. Create Heijunka 5’s Matrix to Control the Dynamic Gemba as Lean Tool. Journal of Human Resource Management. Vol. 3, No. 2, 2015, pp. 6-16. doi: 10.11648/j.jhrm.20150302.11

Abstract: The heijunka tracking matrix has a direct impact on turn overtimes of the inventory to revamp existing orders as well as new technique, and increase the utilization of the inventory’ labors by eliminating the waste in transportation and over-processing actions. The tracking problem is interested in rearranging the location of VSM (e.g., items, pallets and stations) in a grouping of cellular aspect using 5’S concepts, which increase performance level. One of the successful recommendations used in this context is traveler salesman technique but with periodically update (i.e., Dynamic Gemba), it may be need reformulation. Its importance is the capture of work, information, and material (WIP, setup time, process time/unit, error rates, idle time, etc.) which is essential in quantifying and determining waste in terms of cost, delivery time and transportation frequency. This paper is interest in review the dynamic Gemba model to improve the inventory handling via modifying the traveler salesman model with aided proposed heuristic procedures, which are interested in arrange the inventory in minimum time and maximum flexibility in preparing the different orders. The Gemba tasks displayed via (Gemba KPI/DOE board) which contain cost and time information (NNVA: preparing tasks, VA: traveling||tracking tasks and NNVA: unloading tasks). The optimized cost and time analysis after reducing NNVA time and total VA time were sent to the central heijunka matrix. The methodology which includes elimination movements, motion and extra actions wastes are lean, and aim to increase the ability of prepare a lot of service in minimum time.

Keywords: Lean, Manufacturing Facility Layout, Gemba, Heijunka Cellular Manufacturing, KPI in Six-sigma

1. Introduction

The Gemba is a Japanese word may use in different sentence such as Japanese detectives call the crime scene Gemba, and Japanese TV reporters may refer to themselves as reporting from Gemba. In business, Gemba refers to the place where value is created; in manufacturing the Gemba is the factory floor. It can be any "site" such as a construction site, sales floor or where the service provider interacts directly with the customer [1]. In lean manufacturing, the idea of Gemba is that the problems are visible, and the best improvement ideas will come from going to the Gemba. The Gemba may be static or dynamic. The dynamic Gemba changes every visit. The Gemba walk, much like Management By Walking Around (MBWA) to study the best procedures increase performance of executing activities that takes management to the front lines to look for waste and opportunities to practice Gemba heijunka, or practical shop floor improvement. A heijunka is a concept originally created by Toyota for achieving a smoother flow. This paper interest in creating (The heijunka DOE board matrix) which is generally a board schedule which is divided into a grid of KPI equations. Each variable of KPI representing a specific level for significant variable impact on increasing stores performance. The KPI equation contain some Colored (green: executed at the first, yellow: ready to execute, red: postpone to end of track) positions representing individual objective are placed on the heijunka board to provide a visual representation of the good track. The heijunka board make it easy to see what is the smooth track of tasks are queued for completing and for when they are scheduled.

Gemba walks denote the action of going to see the actual process, understand the work, ask questions, and learn [11][12]. It is known as one fundamental part of Lean management philosophy [13].

Gemba Walk is designed to allow leaders to identify existing safety hazards, observe human resources and equipment conditions, ask about the practiced standards, gain knowledge about the work status and build relationships with employees [14]. It is different from classic management style of MBWA. The objective of Gemba Walk is to understand the value stream and its problems rather than review results or make superficial comments [15]. A long with Genchi Genbutsu or "Go, Look, See", Gemba Walk is one of the 5Lean guiding principles that should be practiced by Lean leaders on a daily basis. The Gemba walk is an activity that takes management to the front lines to look for waste and opportunities to practice Gemba kaizen, or practical shop-floor improvement [16].

Dynamic Gemba may be classified an application of group technology (GT), utilizes the concept of divide and conquer and involves the grouping of resources into cells responsible for manufacturing or assembly of similar parts or products [8]. The main reason that stores are attracted towards implementing CM layout is that the benefits of CM and Layout designs, which can normally be realized with relatively low capital investment by relocating and possibly duplicating certain resource as opposed to other automated strategies.

The design for cellular manufacturing involves three stages (i) grouping of items and production equipment into cells, (ii) allocation of the orders cells to areas within the shop floor (inter-cell or facility layout), and (iii) layout of the machines within each cell (intra-cell or machine layout)[16].

Although the concept of generally out problem has been studied for decades [10] for a comprehensive review of such models, certain specific and practical constraints normally present in a real CM environment restrict the implementation of general-purpose layout models [6].

A major difficulty with the existing CM models is that they has no leveling and consider the travelling cost as the sole criterion and the layout with the minimum cost is presented to the decision maker (DM), ignoring many real-world constraints. Because of these limitations many industrial layout designs are being developed by the companies based on their in-house expertise.

This paper demonstrates the capability of handling the inventory via the heijunka DOE board matrix which depends on dynamic Gemba for one of bathtub manufacturing company in Egypt. Furthermore, the capability of the proposed model in capturing realistic constraints such as closeness relationships, location restrictions/preferences, orientations, aisle structures and irregular site shapes is also presented. A facility layout is an arrangement of everything needed for production of goods or delivery of services. A facility is an entity that facilitates the performance of any job.

2. The Track of Dynamic Gemba

This section presents the track of the dynamic Gemba which impact on delivery time via the heijunka a KPI/DOE board matrix. "Fig.1" reviews the formulation of the Gemba and the position of proposed heuristic procedures of 5’S in treating traveler sales man problem in heijunka layout, which has specific property as rapid modification.

Fig. 1. The tree of the different factors for the dynamic Gemba VSM.

"Fig.1"illustrates a possible rough tree representation of the different factors for the dynamic Gemba layout and its related constraints. The proposed model is compared with its validate with the current state (as-is), this figure present a summary for the problem scope.

3. Proposed Heuristic Model

The proposed algorithm deal with movable (e.g. items) and fixed (e.g. cities) stations. The two types of stations have shared in the same objectives, which are reducing total delivery time and costs. The proposed procedures are divided into two phases, the first interested in rearrange the movable items in multi-floors depend on strongly KPI relationship using the heijunka matrix. While reducing the travelling time between the fixed stations in according to analyze the KPI of the delivery time after fed in the heijunka matrix is aim of the second phase.

A. Phase-1: Construct the Gemba Cellular Arrangement

This phase is concerned with reducing the buffers of the items between the different stations that appeared in VSM via grouping heijunka cells.

1.          Determine the KPI for neighboring Hamiltonian stations

2.          Construct the DOE for the KPI significant factors

3.          Feeding the heijunka matrix with KPI values

4.          Sum the values for every row and column.

5.          Choose the largest value, whether from row or column for every station (item||city).

6.          Arrange these values in ascending order.

7.          Compilation of two smaller numbers, and then put the integrated station (item||city) in its position of the ascending order as done in step-3.

8.          Continue until there remains not one single station (item||city), or the collected values become greater than the larger summation of all collected numbers that computed in step-1, whether for row or column.

9.          Determine the cells as appeared in step-5.

10.       Construct the Cells' table and its related station (item||city).

11.       Arrange the related station (item||city) with respect to its strong KPI.

12.       When we put the station (item||city) into place, examine the row for the executed and the extent of his KPI with other already existing, and then the level is chosen depending on the intensity of the KPI, either ground level or upper level.

13.       Continue until put all station (item||city) in the 3-D studio dynamic Gemba.

14.       These attempts depend on heuristic procedures as numbered in phase-1. If the utilization of the station "" and a Traveling Time for any iteration is greater than next iteration, then go on, but if them equal then stop and become known.

4. The Total Cost Formulation

The objective functions studying the cost of traveling between movable or fixed stations. These functions depend on some sets such as the item type(i), intermediate buffers(b),the holding cost that appeared when stations late to pull item to serving order (h), supermarket capacity(sm), tact time(t), also have two parametersis the preparing time for item i on a specific station andwhich denote to unit transferee time for product i via supermarkets.

Fig. 2. Gemba of the travelling chart.

The intermediate inventories addressed as different stores as illustrate in "Fig.2", these inventories were controlled with respect to the pulling operation. The cost computations were illustrated in four types of obj. function sand some constraints represented from"Eq.1toEq.18".

B. Phase-2: Economic Transportation Distance and Time

The economic distance between different stations depends on the transportation cost and market profit form. If there is any units not handled weather transportation system become in position will increase transportation cost and that is a worst case.

The fixed cost of the transportation system is $4km/day; the cost of handling items during trip is $35km/day above the fixed cost. The historical file illustrates that there is a fluctuation between 250and 500km/trip, if the items handled in its time will save $120/day (orders) but else if not handled in time (late, postponed or canceled) will cost company $160 for transportation and $45for inventory.

The next simulation code used to estimate the optimum quantity to shipping and transported during economic trip to design a suitable super-market:

%number of simulation runs





















plot(p(:,1),p(:,2),'+',p(:,1),p(:,2),'-'), xlabel ('Ideal Distance vs. void fill'), ylabel ('Transportation saves$')

"Fig.3" illustrates the ideal# of cartons save the transportation costis more than 63 carton/trip that represents the minimum void for the pace maker quantity.

Fig. 3. The ideal tripdistance vs. void sipping cost.

The Abbreviations used in the objective functions and proposed constraints:

Parameters Decision variables

Order preparing time for station i from station m

Production quantity of product i produced from m in min

Order preparing time for station i in supermarket sm

Available time of supermarket in min.

Available time of labor in min.

Over time capacity used in supermarket sm in min.

Overtime capacity of order i in period t.

Over time capacity of product i stay on station m in min.

Overtime capacity consumed in sm.

Inventory of product i produced from m in min

Regular time working capacity in man-hours

Amount of deficit of product i in period t in min

Overtime working capacity in man-hours

Amount of product i transported from m to sm

Total number of workers to be assigned to sm

Amount of product i transported from sm to m

The quantity of items i demanded by sm

Workforce assigned to pull supermarket

Product holding capacity of sm as raw items

Amount of end of period sm of product i in min

Product holding capacity of sm as semi-finish

Amount of deficit of product i at sm in period t

Transportation time from m to sm in transshipment

Amount of sm of product i in the end

The available distribution time of sm in transshipment

The number of travel needed from m to sm in transshipment

Transportation time from m to sm in pure source

The number of travel needed from m to sm in pure mach.

The available distribution time,

The number of travel needed from m to sm in transshipment

Loading capacity of a Freight car, per travel

Amount of product i transported from sm to next sm

Unit variable production cost of product i, on m

Cost of the Fright car / hrs

Cost of supermarket, for product i from m

Cost of the supermarket in transshipment sectors

Transportation cost

Cost of holding Unit from product i, in sm

The cost of the Overtime for m

Cost of shortage product i, if reduced in sm

Cost of available regular time

Cost of traveling needed i from sm to m

Ÿ The objective functions:

Objective-1, Profit:

Objective-2, Delivery Cost:

Objective-3, Super-Market Cost in transshipment sector:

Objective-4, Super-Market Cost in pure sources or destinations sectors:

Ÿ The related constraints:




















5. Construct the DOE for the KPI

The DOE depends on KPI (delivery time) = f ( preparing order time, travelling time, unloading time ) *f* handling factor. All values is divided by10. The frequency value means the staying time in specific node, the standard staying time have a value (1) or its multiplication.

In our problem we needed to specify the delivery time for every node and number of strokes (f) for every node on the map.

Table 1. The doe of thekpitest.

Run number A (min) (10 : 20) B(min) (50 : 240) C(min) (15 : 60) AB AC BC ABC Y [KPI: delivery time in min(four replicates, n=4] Total
preparing time: X1 travelling time: X2 unloading time: X3         1 2 3 4
1 -1 -1 -1 +1 +1 +1 -1 18.2 18.9 12.9 14.4 64.4
2 +1 -1 +1 -1 +1 -1 -1 27.2 24.0 22.4 22.5 96.1
3 -1 +1 +1 -1 -1 +1 -1 15.9 14.5 15.1 14.2 59.7
4 +1 +1 -1 +1 -1 -1 -1 41.0 43.9 36.3 39.9 161.1
summation 381.3
Effect of every variable on the target  
Sum of square for every variable (SS)  
Delivery time = 23.83 + 8.315X1+3.77X2-4.355X3+4.355X1X2-23.83X1X2X3 23.83

Table 1. Continue.

Total contract
64.4 -64.4 -64.4 -64.4 64.4 -64.4
96.1 96.1 -96.1 96.1 -96.1 -96.1
59.7 -59.7 59.7 59.7 -59.7 -59.7
161.1 161.1 161.1 -161.1 161.1 -161.1
381.3 133.1 60.3 -69.7 69.7 -381.3
  16.63 7.54 -8.71 8.71 -47.66
  1107.22 227.25 303.6 303.6 9086.85
23.83 8.315 3.77 -4.355 4.355 -23.83

"Table1" illustrates the Effect of every variable is equal towhere the N is the total number of runs. The main effect to fA iswhich the average response is for A at high level minus the average of response for A at low level. Therefore, the A=,Sum of squares (SS) is the basis for the analysis of variance computation; the formula for the sum of squares isTo complete heijunka a matrix, we also need SST and SSE. In two-level factorial design., therefore theand the sum of square error is SSE=SST-SSA-SSB-SSC-SSAB-SSABC,SSE=1709.83-1107.22-227.25-303.6-303.6-9086.85=-9318.69

Then the KPI of the heijunka a matrix will be adjusted according to this equation to calculate the KPI from node(i) to node(j), Delivery timeitoj=23.83+ 8.315X1+ 3.77X2-4.355X3+4.355X1X2-23.83X1X2X3. "Fig.5" illustrates the ideal time for preparing orders is 1.6464 hrs, if the labors work 5 hrs after finishing their loading tasks, this means they can prepare between 39:54 orders/day. The best utilization for the drivers comes from working 8.0182hrs/trip plus unloading time to become 9.5182hrs.

Fig. 4. The significant factor of KPI.

Fig. 5. The optimum values of KPI factors.

6. Applying the Proposed Algorithm

"Table2" states the logistics frequency between different stations. Whereis number of trips moved from station(i) to station(j), symbol A is preserved to pure source node. The numbers appeared in "Table2" illustrate the KPI (delivery time)=f(preparing order time, travelling time, unloading time)*f*handling factor. The frequency is considers, how many time will stay in this node. Any negative value is refused. The maximum values indicate the preferred place to start distribution from it. While the minimum values indicate the cells of cities (stations) that increase the efficiency of the delivery action.

Fig. 6. The spaghetti diagram of "as-is".

Table 2. Heijunka tracking matrix: The delivery time among stations ||nodes||cities.

Direction of trip Bypass between stations(مشوارالذهاب)  
S1 S2 S3 S4 S5 S6 S7 S8 S9
backtracking S1   40 70 38   50 63   12 273
S2 6       28 39 41   17 131
S3       37 29 18 19     103
S4     37     21   45 10 113
S5   6 29     33 33   33 134
S6   5 18       46   10 79
S7   17       46   73 73 209
S8       45     21     66
S9 12 17         73     102
  18 85 154 120 57 207 296 118 155  

Transporting Cost(as-is)==£756.2. The travelling distance for "Fig.6" is 470km for single trip; this travelling is consumed 470*250day/year=117,500km/year.

1.   Sum the relationships for every row and column.

2.   Choose the smallest value, whether from row or column for every node and arrange these values in ascending order.

  S8 S5 S2 S4 S9 S3 S6 S1 S7
row 66 134 131 113 102 103 79 273 209
column 118 57 85 120 155 154 207 18 296

184 191 216 233 257 257 286 291 505

1.   Compilation of two smaller numbers, and then put the integrated nodes (cities in the same trip) in its position of the ascending order as done in step-2.

iteration S2 S4 S9 S3 S6 S1 S8S5 S7
1 216 233 257 257 286 291 375 505
  S9 S3 S6 S1 S8S5 S2S4 S7
2 257 257 286 291 375 449 505
  S6 S1 S5S8 S4S2 S7 S3S9
3 286 291 375 449 505 514
  S5S8 S4S2 S7 S3S9 S6S1
4 375 449 505 514 577
  S7 S3S9 S6S1 S5S8S4S2
5 505 514 577 824
Cell S3S9 S6S1 S7S5S8S4S2
  514>505 577 1019

2.   Continue until there remains not one single node, or the first cell for the collected values become greater than the larger summation of the end value for the delivery time for this city, whether for row or column. This number in this case is 505.

3.   Determine the cells as appeared in step-4.

Cells S3S9 S6S1 S7S5S8S4S2

4.   Construct the Cells' table and its related nodes.

Cell_1 Related nodes||cities
S3 S4 S5 S7 S6  
S9 S7 S5 S2 S1
Cell_2 Related nodes||cities
S6 S1 S7 S2 S5 S4
S1 S3 S7 S6 S2 S4
Cell_3 Related nodes||cities
S7 S9 S1 S6 S2 S8
S5 S7 S6 S3 S2  
S8 S7 S4      
S4 S8 S1 S3    
S2 S7 S6  

5.   Arrange the related nodes with respect to its strong delivery relations, and construct the multi-floor trip, which depends on TPS. Which suggest to build the store in S7 position,

Fig. 7. New 3-D Gemba stations "Should-be".

Transporting Cost (heijunka) = 80,000 km/year, these procedures which present "Fig.7" were save 30% of costs. "Table3" illustrates the developing occurred by applying proposed procedures than current as-is situation.

Table 3. The Comparison between "proposed algorithm" and "as-is" state.

Attributes As-is To-be Developing%
Delivery Lead Time "min" 713 595 16.5%
Travelling distance for trip km/year 117,500 80,000 32%
Travelling cost for complete trip£ 645.6 482.6 30%
Fatigue and strength High Very Low  
Utilization to meet target X Up 28%
Inter-cell material handling turns Complex good  
Trip utilization Low enhance  
Number of cells None deterministic  
Quality feedback Days Minutes  
The Profits relate void cost 1227123 2001983 63%
The Delivery Costs 315,234 147,235 53%

7. The Conclusion

The dynamic Gemba VSM procedures are affected by the delivery time and costs. The proposed procedures are compared with as-is situation and aid the TSP to enhance the solution. The Delivery lead time was developed by 16.5%, the travelling time (NNVA) was developed by 32% and the travelling cost was reduced in average by30%. The proposed procedures present a cellular of items, which increase from the labor utilization and quality feedback. Also increase the profits by 63%, all this information was illustrated in"Table3".


  1. Aiello G, Enea M, Galante G,. Multi-objective approach to facility layout problem by genetic search algorithm and Electre method. Robotics and Computer-Integrated Manufacturing, 22,447–455.(2006)
  2. Aleisha EE, Lin L. For effectiveness facilities planning: Layout optimization then simulation, or vice versa? In Proceedings of the (2005).
  3. Asef-Vaziri A, Laporte G. Loop based facility planning and material handling. European Journal of Operational Research, 164 (1), 1–11,(2005).
  4. Bazargan-Lari M., Kaebernick H. Intra-cell and inter-cell layout designs for cellular manufacturing. International Journal of Industrial Engineering ± Applications and Practice 3,139±150,(1996).
  5. Bazargan-Lari M., Kaebernick H. An approach to the machine layout problem in a cellular manufacturing environment. Production Planning and Control8(1),41±55,(1997).
  6. Baykasoglu, A., Dereli, T., Sabuncu,I. An ant colony algorithm for solving budget constrained and unconstrained dynamic facility layout problems. Omega, 34(4),385–396,(2006).
  7. VanCamp, D.J., Carter, M.W., and Vannelli, A. Anon-linear optimization approach for solving facility layout problems. European Journal of Operational Research57,174±189,(1992).
  8. Heragu, S.S.Group technology and cellular manufacturing. IEEE Transactions on Systems, Man, and Cybernetics 24(2), (1994).
  9. Heragu, S.S. and Kusiak, A. Efficient models for the facility layout problem. European Journal of Operational Research 53,1±13,1991.
  10. Jumpup ^Womack, Jim. Gemba Walks. Lean Enterprise Institute, Inc.p.348.ISBN978-1-934109-15-1,(2011).
  11. Jumpup^
  12. Jumpup ^
  13. Jumpup ^Andrew Castle, Rachel Harvey. "Lean information management: the use of observational data in healthcare".InternationalJournalofProductivityandPerformanceManagement,Vol.58Iss:3,pp.280-299.ISSN1741-0401.,(2009)
  14. Jumpup ^Mann, David W. (2005). Creating a lean culture; tools to sustain lean conversions. Productivity Press. p. 211.ISBN1563273225.
  15. H. Ku, H. Wang, N. A solution to the unequal is a facilities layout problem by genetic algorithm. Computers in Industry, 56(2), 207–220,(2005).

Article Tools
Follow on us
Science Publishing Group
NEW YORK, NY 10018
Tel: (001)347-688-8931