Engineering, Technology & Applied Science Research Vol. 8, No. 4, 2018, 3103-3107 3103 www.etasr.com Odedairo & Nwabuokie: Framework for Operational Performance Measurements in Small and Medium… Framework for Operational Performance Measurements in Small and Medium Scale Industries Using Discrete Event Simulation Approach Babatunde Omoniyi Odedairo Department of Industrial and Production Engineering University of Ibadan Ibadan, Nigeria bo.odedairo@ui.edu.ng Nnamdi Nwabuokie Department of Industrial and Production Engineering University of Ibadan Ibadan, Nigeria valiantnnamdi@yahoo.ca Abstract—Globally, production systems must cope with limitations arising from variabilities and complexities due to globalization and technological advancements. To survive in spite of these challenges, critical process measures need to be closely monitored to ensure improved system performance. For production managers, the availability of accurate measurements which depict the status of production activities in real time is desired. This study is designed to develop an operational data decision support tool (ODATA-DST) using discrete event simulation approach. The work-in-process and processing time of each workstation/buffer station in a bottled water production system were investigated. The status of each job as they move through the system was used to simulate a routing matrix. The production output data for 50cl and 75cl product from 2014-2016 were collected. A mathematical model for routing jobs from the point of arrival to the point of departure was developed using discrete event simulation. A graphical user interface (GUI) was designed based on the factory’s performance measurement algorithm. Simulating the factory’s work-in-process with respect to internal benchmarks yielded a cycle time of 4.4, 6.23, 5.04 and throughput of 0.645, 0.455, 0.637 for best case scenario, worst case scenario and practical worst case scenario respectively. The factory performed below the simulated benchmark at 26%, 28%, 28% for the 50cl and at 51%, 54%, 59% for 75cl regarding the year 2014, 2015 and 2017 respectively. Performance measurement decision support tool has been developed to enhance the production manager’s decision making capability. The tool can improve production data analysis and performance predictions. Keywords-performance measures; production system; discrete event simulation;decision support system I. INTRODUCTION The need for continuous performance improvement in a production system despite the complexities arising from market fluctuations will continue to drive the desire for innovative research. Performance measurement, a sub-division of performance evaluation involves the selection of appropriate quantitative measures to aid decision making in a system. These measures are vital input into any decision support tools (DST) [1, 2]. Also, such measures are required to assist executives at different decision levels. Ultimately, these decisions will contribute to the actualization of the strategic goals of the organization [3-7]. Authors in [8] classified performance measures into: (1) measure focus, and (2) measure tense. The former comprises of financial (monetary) and operational (non-monetary) data, while measure tense entails studying the past to improve the present. Diverse studies have been carried out on how to measure system performance using DST [6, 9-11]. This is necessary as the profitability, productivity, and survivability of any production system largely depend on the quality of the decisions obtained from such tools. DST in a production system can be deployed at operational, tactical and strategic levels. However, due to the ambiguities associated with most decision processes, the need to smoothen the complexities associated with choosing the best alternative cannot be ignored [12]. Decision support tools relevant to the production system include the following: (1) thermodynamics and exergy analysis, (2) optimization, and (3) simulation [13-16]. On simulation, the aim is to imitate real-world process over time [17, 18]. Also, in a simulation model, discrete mathematics can be employed in which events of various kinds are kept and governed in a queue for each object [19]. Discrete event simulation (DES) considers state changes at discrete points (points which an event occurs). It can be used to answer “what if” scenarios, diagnose the occurrence of certain phenomena and enhance system development over time [12, 17, 20-22]. Despite the increase in the research work on using DES as a DST, empirical studies have shown that it is minimally used in production systems [1, 12, 23-25]. In Nigeria, one of several challenges limiting the performance of small and medium scale enterprises (SMEs) involved in production process is the lack of access to proprietary DST [7, 26, 27]. Based on this reality, in this study the objective is to develop an operational data DST (ODATA- DST) for a bottled water factory using DES analytical approach. The rest of this paper is structured as follows: A brief discussion on DST and DES is the focus of section II. In section III, ODATA-DST was developed using an illustrative example. Results from the example and conclusion are the focus of sections IV and V respectively. pec lim loa bot qu pro im a ind exa mo inc ava lin vag dec ma (2) Mo ma pro (2) pre ma Di sim per to set ins res dis pro bas aut for aut sys opt eff per op beh A. pla pro II. fol dis pro bec eve Engineerin www.etasr In [18] auth culiar to most mited to, the fo ad resources, ttlenecks and euing and del ocess and (7 mprove efficien performance dustries usin amination. Au odel to measur creasing comp ailable to org nguistic variab gueness and u cision making aking process ) documentatio odels are used any problems. ocesses in mo ) state estim ediction, (6) athematical im screte event mulation meth rformance of abstract the dy t of discrete stants of time sources and is screpancies in oduction syste sed real-time tomation syste r planning, pr thors utilized stem design. timization to ficiency indust The mathem rformance me erational perf havior of the p Descriptive P The flow pro ant is describ oduction line The worksta llowing the screte event oduction line cause their ac ent, there is a ng, Technology r.com II. RELAT hors suggeste t production sy ollowing: (1) , (2) labor d choke poin lays caused by ) scheduling ncy of the supp measurement ng data ca uthor in [10] d re business pe plexity and h ganizations. H bles have the uncertainty nat g. Authors in [ is driven by th on, (3) knowle d to demystify . Authors in odeling as: (1) mation, (3) m hypothesis mage to a pro simulation hod universa a production ynamic and sto sequence of e [30]. Also s useful in sol simulated env ems, authors decision ma em. Authors i roduction and d DES for c Author in o improve d try. III. ME matical model p easurement wa formance mea production line Problem Defin ocess of a sma bed in Figure with 9 workst ations are arr same process simulation will be limi ctivities are d change in the y & Applied Sci Odedairo & N TED LITERATU ed some perf ystems, these throughput un and machin nts, (4) staff y material han effectiveness ply chain, auth t system for ategorization, developed a fu erformance as eterogeneous He concluded t ability to dea ture of produc [13] comment he following: edge and (4) m y the complex [28, 29] iden ) data assessm monitoring, ( generation, oblem, and (8 (DES) mode ally suitable system. The m ochastic behav events by co , DES dema lving statistica vironments. O in [31] deve aking tool fo in [32] presen d scheduling customer driv [19] combine decision supp ETHODOLOGY proposed in [3 as adopted. In asures required e are defined. nition all scaled bottle e 1. The pro tations (A1-A ranged in ser sing sequence framework, t ited to works discrete events e state of the p ience Research Nwabuokie: Fr URE formance mea include, but a nder mean and ne utilization f requirements ndling, (6) wo s. On the ne hor in [11] des r Thai autom clustering, uzzy data ware a result of the nature of the that fuzzy set al with imprec ction data for ted that the de (1) communic model develop xity associated ntified some o ment and redu (4) diagnosis (7) creatio 8) decision ma el is a com for modeling method can be vior of a system onsidering eve ands less com al uncertaintie n the use of D eloped a simu or a manufac nted the use of decisions. In ven manufac ed simulation ort in an e 34, 35] on fac Table I, the v d to understan ed water produ ocess consists A9) defined in ries with eac e. In design the study o stations A2 t s. For each di product as it m h V ramework for O asures are not d peak n, (3) s, (5) ork in eed to signed motive and ehouse e ever- e data ts and cision, better ecision cation, pment. d with of the uction, s, (5) on of aking. mputer g the e used m as a ent in mputer es and DES in ulation cturing f DES [33], cturing n and energy ctory’s arious nd the uction of a Table ch job ing a of the to A9 iscrete moves from pro wor in F B.        T C. foll  Vol. 8, No. 4, 20 Operational Perf m one workst oduct. Each w rkstation. Job Figure 2. Fig. 1. P Model Assum The bottled w finished good The cycle ti throughput tim been empty o Processing ti varies from on Buffer capaci The product exponential d Blockage of a buffers are be The factory downtimes, n constantly in Notation WIP CT CTbest Tapprox Capprox. TH CTworst PR W0 THbest Ma N W To T THworst Mathematica To model O lowing measur Cycle time Case 1: When 018, 3103-3107 rformance Mea tation to anot workstation (A routing of the Plant layout of the mptions water factory ds buffers. ime of the m me varies with r saturated). me in each w ne workstation ity between su tion line is distribution fail a station occu eyond its capac production no loss in pro operation. TABLE I. W Cycle tim Time for a job to C Cycle time Criti aximum throughp N Leve R Ave Throughput for l Model ODATA-DST res were adopt n the productio 7 asurements in S ther until it b A2-A9) has e production s e bottled water pro is a serial pr machines is c h respect to it workstation is n to the other. uccessive work reliable with lure rate. urs if the jobs city. line model oduction due BASIC NOTATION Definition Work–in–process Cycle time me at best case per o go through an u Capacity of the lin Throughput for worst case pe Production rate ical Work–in–pro ut at best–case pe Number of station el of work–in–pro Raw process time erage processing t worst case perfor Bottleneck rate T using DES ted. on line is relati 3104 Small and Medi becomes a fin at least one system is pres oduction plant. roduction line constant while ts state (in term s deterministic k stations is fin h a steady at the downst has no ma to waste and NS s formance uncongested line ne erformance cess erformance scenar s ocess e time rmance scenario S framework, ively empty ium… nished sub- ented with e the ms of c but nite. state tream chine d it is rio , the Engineering, Technology & Applied Science Research Vol. 8, No. 4, 2018, 3103-3107 3105 www.etasr.com Odedairo & Nwabuokie: Framework for Operational Performance Measurements in Small and Medium… = (1) Case 2: When the production line is saturated = . , . (2)  Machine Capacity . = (3)  Throughput = ∗ (4)  Work-in-process (from Little’s Law) = × (5)  Bottleneck rate = (6)  Average time at a station = (7) TABLE II. PLANT NAMING CONVENTION Conventions Name Description A1 Belt conveyor Belt conveyor A2 Buffer station 1 Buffer station 1 A3 Workstation 1 Automatic rinsing machine A4 Buffer station 2 Buffer station 2 A5 Workstation 2 Automatic filling machine A6 Buffer station 3 Buffer station 3 A7 Workstation 3 Automatic capping machine A8 Buffer station 4 Buffer station 4 A9 Workstation 4 Automatic Packaging machine A10 Warehouse Warehouse D. ODATA-DST Graphic User Interface The computer implementation of ODATA-DST was achieved using visual basic application (VBA) integrated development environment. The following motivated the use of VBA: (1) compatibility, (2) availability across multiple platforms, (3) interactive nature, and (4) ease of numerical programming. A screenshot of ODATA is shown in Figure 3. E. Best-Case Scenario, Worst-Case Scenario and Pratical Worst- Case Scenario of the Production Line Analyzing the production line based on the best-case scenario, worst-case scenario and practical worst-case scenario is essential. These parameters are required to measure performance, determine the possible behavior, and areas requiring improvement at any time period. 1) Best-Case Performance (BCP) This can be classified into minimum cycle time and maximum throughput. Minimum cycle time for a given WIP level (w) is given by: = ≤ _0 ℎ (8) Maximum throughput for a WIP level given by : = ≤ ℎ (9) START Arrival and Movement to Workstation 1 Workstation 1 Idle Buffer 2 Process job in workstation 1 Buffer Next workstation No Yes Yes Yes No Yes Yes Fig. 2. Flow chart of job routing. 2) Worst-Case Performance (WCP) This involves the maximum cycle time and minimum throughput possible for a line with bottleneck rate and raw process time ( ). Equation (10) is the worst-case cycle time for a given WIP level w. = (10) Equations (11)-(13) are the worst-case throughput for a given WIP level w. = (11) = 1 + (12) = + (13) Applying Little’s law, the corresponding throughput will be: = (14) = (15) 3) Practical Worst-Case (PWC) Equation (16) is the cycle time for a PWC. = + (16) bot req tim 2.2 pro bes rel Fo out No out rel and Engineerin www.etasr The PWC thr= The results a ttleneck of 0 quired to achi me of 2.2 minu 2 minutes. Th ocess time, as st-case, worst lation to cycle or 50cl bottled tput were 16 ovember respe tputs were 28 lationships bet d 75cl are show Station Nu Buffer WorkStat Buffer WorkStat Buffer WorkStat Buffer Work Sta Av W (Bottles) 10 20 30 40 60 70 ng, Technology r.com roughput for a Fig. 3. ODA IV. are presented i 0.645jobs/minu ieve maximum utes is 1.419 c his is lower th proposed in [ t-case and pr e time and thro d water produ 6% and 40% ectively. For th % and 83% in tween work-in wn in Figures umber Numb r 1 tion 1 r 2 tion 2 r 3 tion 3 r 4 tion 4 verage Processin Bottleneck W (Cartons) 0.8 1.7 2.5 3.3 5.0 5.8 y & Applied Sci Odedairo & N a given WIP le ATA – DST scree RESULTS in Table III. W ute, the critic m throughput artons. The ra han the sum o 34]. The relati ractical worst oughput is pre uction, minim % and occurr he 75cl, minim n July and Ma n-process and 4 and 5 respe ber of Machines 4 4 4 4 4 3 4 1 ng time Best C CTbest (mins 2.2 2.6 3.9 5.2 7.8 9.0 ience Research Nwabuokie: Fr evel is given by (17 en Workstation 4 cal work-in-pr at raw proce aw process tim of each works ionship betwe t case scenari esented in Tab mum and max red in March mum and max ay respectively throughput fo ctively. TABLE III. Process Time 3 8 4 8 5 3 21 93 18.125 0.645 TABLE IV. Case Scenario s) THbest (m 0.379 0.645 0.645 0.645 0.645 0.645 h V ramework for O y: 7) has a rocess essing me was station en the ios in ble IV. ximum h and ximum y. The or 50cl dev cap mac pro too per Fig. Fig. RESULTS FRO es (secs) Proce 5 5 INTERNAL BEN W mins) CTwors 1.8 3.6 5.5 7.3 11. 12. Vol. 8, No. 4, 20 Operational Perf A performan veloped to enh pability. Perfor chine capacity ocess time we l is capable formance pred . 4. Relations . 5. Relations OM ODATA-DST ess Times (mins) 0.050 0.133 0.067 0.133 0.083 0.050 0.350 1.550 Raw Proces Critical W NCHMARK OUTPUT Worst Case Scen st (mins) THw 83 67 50 33 .00 .83 018, 3103-3107 rformance Mea V. CO nce measurem hance product rmance measu y, work–in–pr re used to de of improving dictions. hip between throu hip between throu ) Jobs/Minute 80 30 60 30 48 60 11.428 0.645 ss time WIP T nario worst (mins) 0.455 0.455 0.455 0.455 0.455 0.455 7 asurements in S ONCLUSIONS ment suppor tion manager’ ures like cycle rocess, bottlen erive a suitabl g production ughput and work ughput and work e Station Cap 1.3 0.5 1.0 0.5 0. 1.0 0. 0.0 2 1 Practical Wors CTpwc (mins) 1.94 3.23 4.53 5.82 8.40 9.69 3106 Small and Medi rt tool has ’s decision ma e time, throug neck rate and le benchmark. data analysis in process (50cl b in process (75cl b pacity (Job/sec) 3333 5000 0000 5000 8000 0000 1905 0108 2.20 .419 st Case Scenario THpwc (mins 0.619 0.632 0.636 0.638 0.641 0.641 ium… been aking ghput, d raw . The s and bottle) bottle) o s) Engineering, Technology & Applied Science Research Vol. 8, No. 4, 2018, 3103-3107 3107 www.etasr.com Odedairo & Nwabuokie: Framework for Operational Performance Measurements in Small and Medium… REFERENCES [1] J. Heilala, J. Montonen, P. Jarvinen, S. Kivikunnas, “Decision Support Using Simulation for Customer-driven Manufacturing System Design and Operational Planning”, in: Advances in Decision Support Systems, pp. 235-260, IntechOpen, 2010 [2] M. P. Roessler, J. Reimer, M. Mueller, “Decision Support for Choosing an Appropriate Simulation Method for Dynamic Material Flow Analysis”, Journal of Industrial and Intelligent Information, Vol. 3, No. 4, pp. 337-341, 2015 [3] W. J. Hopp, M. L. Spearman, Factory Physics, Irwin McGraw-Hill, 2000 [4] M. Wilcox, M. Bourne, “Predicting Performance”, Management Decision, Vol. 41, No. 8, pp. 806-816, 2003 [5] A. Neely, M. Gregory, K. Platts, “Performance measurement system design: A literature Review and Research Agenda”, International Journal of Operations an Production Management, Vol. 25, pp. 1228-1263, 2005 [6] R. Sundkvist, Financial benefits of shoop floor productivity improvements, PhD Thesis, Chalmers University of Technology, 2014 [7] B. O. Odedairo, D. S. Ladokun, “Varying Lot -Sizing Models for Optimum Quantity- Determination in Material Requirement Planning System”, Presented at the International Conference of Manufacturing Engineering and Engineering Management, London, UK, July 4-6, 2018 [8] S. Melnyk, D. Stewart, M. Swink, “Metrics and Performance Measurement in Operations Management: Dealing with the Metrics Maze”, Journal of Operations Management, Vol. 22, pp. 209-217, 2014 [9] A. Gunasekaran, C. Patel, R. McGaughey, “A framework for supply chain performance measurement”, International Journal of Production Economics, Vol. 87, No. 3, pp. 333-347, 2004 [10] D. Fasel, Fuzzy Data Warehousing for Performance Measurement- Concept and Implementation, Springer, Switzerland, 2014 [11] T. Taweesak, “Performance Measurement System Design and Implementation in Thai Automotive Industry”, PhD Thesis, University of South Australia, 2005 [12] J. Lorentzon, Fredlund, Application of Performance Measurement on Manufacturing Simulations for Knowledge-Based Decision Support, MSc Thesis, Lund University, 2017 [13] P. Sandberg, M. Larsson, J. Dahl, M. Soderstrom, H. Vourinen, “In search of stability-Investigating Flexible and Stable Production Strategies for an Optimised Steel Plant”, International Journal of Green Energy, Vol. 3, No. 2, pp. 159-170, 2006 [14] M. Gong, Using Exergy and Optimization Models to Improve Industrial Energy Systems Towards Sustainablity, PhD Thesis, Linkopings Universitet, 2004 [15] J. Persson, T. Berntsson, “Influence of Short-term Variations on Energy- Saving Opportunities in a pulp Mill”, Journal of Cleaner Production, Vol. 18, No. 9, pp. 935-943, 2010 [16] J. Jonsson, P. Ruohonen, G. Michel, T. Berntsson, “The potential for Steam Savings and Implementation of Different Biorefinery Concepts in Scandinavian Integrated TMP and Paper Mills”, Applied Thermal Engineering, Vol. 31, No. 13, pp. 2107-2114, 2011 [17] S. Robinson, Successful Simulation: A Practical Approach to Simulation Projects, McGraw -Hill, 1994 [18] J. Banks, J. Carson, B. Nelson, D. Nicol, Discrete-Event System Simulation, Third Edition, Prentice–Hall, 2001 [19] N. Mardan, Combining Simulation and Optimization for Improved Decision Support and Energy Efficiency in Industry, Linkoping University Institute of Technology, Sweden, 2012 [20] M. Al Durgham, M Barghash, “A generalised framework for simulation -based decision support for manufacturing”, Production Planning and Control, Vol. 19, No. 5, pp.518–534, 2008 [21] S. Kursun, F. Kalaoglu, “Simulation of Production Line Balancing in Apparel Manufacturing”, Fibres in Eastern Europe. Vol. 12, No. 6, pp. 68-71, 2009 [22] B. Johansson, On Virtual Development of Manufacturing Systems: Proposal for a Modular Discrete Event Simulation Methodology, PhD Thesis, Chalmers University of Technology, 2006 [23] A. G. Hirschberg, K. G. Heitmann, “Simulation in German Industry -A survey”, 9th European Simulation Symposium and Exhibition Simulation in Industry, Passau, Germany, pp. 429-433, October 19-22, 1997 [24] L. Holst, G. Bolmsjo, “Simulation Integration in Manufacturing System Development: A study of Japanese industry”, Industrial Management and Data Systems, Vol. 101, No. 7, pp. 339-356, 2001 [25] J. Fowler, O. Rose, “Grand challenges in Modelling and Simulation of Complex Manufacturing Systems”, Simulation, Vol. 80, No. 9, pp. 469- 476, 2004 [26] B .O. Odedairo, D. Bell, “Framework for Introducing and Implementing Value Methods: A Novel Toolkit for Small and Medium Scale Industries in Developing Nations”, International Journal of Basic and Applied Sciences, Vol. 9. No.10, pp. 87-95, 2009 [27] V. O. Oladokun, A. O Olaitan, “Development of a Materials Requirements Planning (MRP) Software”, Pacific Journal of Science and Technology, Vol. 13, No. 1, pp. 351-357, 2012 [28] J. Spanier, “Thoughts about Essential of Mathematical Modelling”, Mathematical Modelling, Vol. 1, pp. 99-108, 1980 [29] A. J. Jakeman, R. A Letcher, J. P. Norton, “Ten Iterative Steps in Development and Evaluation of Environmental Models”, Environmental Modelling and Software, Vol. 2, No. 5, pp. 602-614, 2006 [30] W. Domschke, A. Drexel, Introduction to Operations Research, Springer-Verlag, 2007 [31] H. I. Yoon, W. Shen, “Simulation -based Real -Time Decision Making for Manufacturing Automation Systems: A review”, Manufacturing Technology and Management, Vol. 8, pp. 188–202, 2006 [32] B. Kadar, A. Pfeiffer, L Monostori, “Discrete Event Simulation for Supporting Production and Scheduling Decisions in Digital Factories”, 37th CIRP International Seminar on Manufacturing Systems, Digital Enterprise, Production Networks, Budapest, Hungary, pp. 444-448, 2004 [33] J. Heilala, J. Montonen, A. Salmela, P. Jarvenpaa, “Modelling and Simulation for Customer Driven Manufacturing System Design and Operations Planning”, IEEE 2007 Winter Simulation Conference, Miami, USA, December 7-10, 2007 [34] W. J. Hopp, M. L. Spearman, Factory Physics: Foundations of Manufacturing Management, Waveland Press, 2008 [35] G. L. Curry, R. M Feldman, Manufacturing System Modeling and Analysis, Springer, 2007