1. Claudu V. KIFOR, 2. Amela BUCUR, 3. Muhammad Arsalan FAROOQ DEVELOPMENT & IMPLEMENTATION OF SUPPLY CHAIN MANAGEMENT FOCUSING ON PROCUREMENT PROCESSES & SUPPLIERS 1. LUCIAN BLAGA UNIVERSITY SIBIU, RESEARCH CENTER IN QUALITY MANAGEMENT AND ENGINEERING, SIBIU, ROMANIA 2. LUCIAN BLAGA UNIVERSITY OF SIBIU, FACULTY OF SCIENCE, DEPARTMENT OF MATHEMATICS, SIBIU, ROMANIA 3. POLYTECHNIC UNIVERSITY OF MILAN, DEPARTMENT OF INDUSTRIAL ENGINEERING & MANAGEMENT ECONOMICS, ITALY ABSTRACT: Supply chan optmzaton s a modern subject, approached by many specalsts n ther papers [2,6,7,9 and others]. It s the applcaton of processes and tools to ensure the optmal operaton of a manufacturng and dstrbuton supply chan. Ths often nvolves the applcaton of mathematcal modelng technques usng computer software. One of the papers related to supply chan optmzaton was publshed n 2007, by Ghodsypour and O Bren, n an Internatonal Journal of Producton Economcs [8]. The paper focused n optmzng the whole supply chan and make t more effectve and effcent by not just focusng on the buyer s cost but by examnng costs from both the buyer and the seller. In ths paper we formulated and ntroduced a dscounted cost savng strategy to ensure that both actors (buyer and seller) beneft. Secondly, we ncluded an out of stock and excess nventory costs and then formulated and mplemented an optmzed soluton. The model has been solved by applyng mxednteger nonlnear programmng usng LINGO software. Ths paper compares the results wth the orgnal work usng the same example and provdes a wn wn stuaton for both the actor to make the problem feasble even for decentralzed stuaton. KEYWORDS: Supply Chan Optmzaton, Inventory Costs, Integer Non Lnear Programmng, LINGO, decentralzed supply chan INTRODUCTION Supply chan management s the process by whch supplers, partners, and customers plan, mplement and manage the flow of nformaton, servce, and products n a way that mproves busness operatons n terms of speed, aglty, real tme control, or customer response (Manthouet al., 2004). Wth ntensve nternatonalzaton and globalzaton of markets, multple dmensons of operatonal capabltes between supplers, tradng partners, and customers have to be developed for frms to succeed. A key operatonal capablty necessary to meet ths challenge s operatonal nformaton systems that are embedded n a frm s core busness processes and ts nter organzatonal busness transactons (Nurmlaakso, 2008). Producton n the early years was smple, wth sngle flow of products movng from raw materal supplers, to manufacturers and then to markets. Nowadays, shorter product lfecycle and ncreasng demand among all have led to a complcated supply chan. Due to cost pressure and compettve advantages, companes are adoptng globalzaton and outsourcng strateges. Ths also requres an extended supply chan network, hence ncreases the nodes n the system. In addton, many companes have ntroduced lean producton concepts, whch ntend to remove wastes from a supply chan, for nstance, by reducng the number of supplers. Ths helps n smoothng the operatons but t would also create problems f unexpected events happen n a supply chan. The rsng use of nternet helps supply network n sharng nformaton vsblty (Chrstopher and Lee, 2004; Lee, 2002, 2004; Narayanan and Raman, 2004). The producton of hghly dversfed products wth short lfe cycles such as computer parts, fashon clothes and some food tems among many others as well as the remarkably hgh levels of competton pushes the dfferent companes towards the ntegraton of dfferent producton and nventory related decsons. Consequently, companes are realzng the necessty of havng elevated levels of mutual understandng and better collaboraton wth ther customers and supplers alke. To reman compettve, frms can no longer operate as ndvdual and autonomous enttes but rather as an ntegral part of the supply chan. The area of supply chan management (SCM) has ganed a lot of nterest from researchers as well as practtoners n the ndustry. In partcular, the ntegrated sngle vendor sngle buyer problem, also called the jont economc lot szng problem (JELP), has receved a lot of attenton n recent years as t represents the buldng block copyrght FACULTY of ENGINEERING HUNEDOARA, ROMANIA 77
for the wder supply chan. Essentally, the retaler (buyer) observes a determnstc demand and orders lots from the manufacturer (vendor). The vendor satsfes ths downstream demand through manufacturng the requested product n lots, where each produced lot s shpped to the buyer n batches. The problem s to fnd the number of shpments and sze of each batch such that the jont manufacturer and retaler cost s mnmzed. For a vertcally ntegrated supply chan owned partally or jontly by the same company, such coordnated producton shpment polcy provdes valuable nsghts and optmal decsons that lead to global optmzaton. On the other hand, when ndvdual enttes are owned separately, such polcy may not beneft all partes equally as some may encounter an ncrease n ther costs and hence become less eager to depart from ther locally optmzed polces. In such stuatons, sharng those benefts resultng from the coordnated approach becomes a major ssue. By usng effectve ncentve systems such as accountng methods, transfer prcng schemes, quantty dscount, etc., the objectve of each partner can be algned to that of the supply chan as a whole (Ganeshan, 1999; L & O Bren, 1999; Agrawal et al., 2004). LITERATURE REVIEW A supply chan s made up by multple actors, multple flows of tems, nformaton and fnances. Each network node has ts own customers and supplers management strateges, demand arrval process and demand forecast methods, nventory control polces and tems mxture. An nterestng approach n studyng the effects of nventory control polces on supply chan performance s proposed by Tagaras and Vlachos (2001). They consder a perodc revew nventory control polcy workng wth two replenshment modes (regular and emergency). Graves and Wllems (2005) propose a more centrc approach on the entre supply chan. They deal wth supply chan confguratons n terms of supplers, parts, processes and transportaton modes to be selected at each stage tryng to mnmze the total supply chan cost. Note that dfferent supply chans are characterzed by dfferent crtcal parameters so a decson makng tool should provde to managers hgh flexblty n terms of scenaros defnton, crtcal parameters and performance measures selecton. For most companes, provdng the customers wth a better servce at a reduced cost s one of the ultmate strategc goals. Most of the work related to JELP has been conducted n the context of a two layer supply chan consstng of a sngle vendor and a sngle buyer. Goyal (1977) suggested a lot for lot polcy wth the assumpton of nfnte producton rate for the manufacturer. Later, Banerjee (1986) mantaned the lotfor lot polcy for the more realstc case of a fnte producton rate. The lot for lot assumpton was relaxed by Goyal (1988) where he assumed that the vendor shps the lot n a number of equal sze shpments. Goyal (1995) developed a polcy where the shpment szes ncrease by a factor ncreasng geometrcally. Hll (1997) generalzed the latter model through consderng the geometrc growth factor as a decson varable. The optmal soluton to the problem n ts general form (.e., wthout any assumptons regardng the shpment polcy) was obtaned by Hll (1999). Goyal and Nebebe (2000) consdered a polcy where the frst shpment s small and the followng shp ments are larger and of equal sze. For a comprehensve revew of the JELP, the reader s referred to Ben Daya et al. (2008). More recently, ths problem has been extended to the case of a three layer supply chan. Khouja (2003) was the frst to consder a three stage supply chan wth one or more frms at each stage. Recently, some researchers have developed jont decson makng n multple suppler sngle buyer supply chans. Km et al. (2005) developed a producton delvery polcy n a supply chan consstng of a sngle manufacturer wth multple plants n parallel and a sngle warehouse, and a sngle retaler. They bult a model to determne an optmal producton cycle length (or nterchangeably producton lot sze) for the manufacturer, a delvery polcy (.e. frequency and quantty) to the retaler and a producton allocaton scheme for multple plants so as to mnmze the average total cost ncurred by both the manufacturer and the retaler. Yung et al. (2006) presented heurstcs solutons to a jont decson problem for a sngle product and for multple products n a producton dstrbuton network system wth multple supplers and multple destnatons. Park et al. (2006) developed a mathematcal model n whch the retaler places orders based on the EOQ polcy and allocates them to the multple manufacturers. In ther model producton, allocaton ratos and shpment frequences at the manufacturers as well as purchasng cycle length at the retaler are formulated to mnmze the average total cost that ncludes average total cost at the manufacturers and retaler. Van Der Veen and Venugopal (2005) ntroduced a revenue sharng model n the vdeo rental supply chan consstng of two ndependent frms, namely a move studo and a vdeo rental shop. They 78 Tome X (Year 2012). Fasccule 2. ISSN 1584 2665
dscussed three scenaros, namely soltare or decentralzed scenaro, partnershp and revenue sharng. Ther paper showed that the decentralzed stuaton could not optmze the chan, but revenue sharng contracts could lead to whole optmzaton of the supply chan wth a wn wn stuaton. Ths paper s the extenson of the work by Ghodsypour and O Bren (2007), the same model has been used wth multple suppler sngle buyer coordnaton model, whch optmzes the total supply chan but n order to create a wn wn stuaton for both the actors (buyer and seller) a dscounted cost savng strategy has been formulated, and n the second part an out of stock and excess nventory costs has been ntroduced. METHOD The objectve functon s to mnmze the total annual cost of a supply chan that ncludes suppler and buyer annual costs, the followng model has been used whch was mplemented by Ghodyspour and O Bren (2007); s.t. X D P, S Mn (ASCT) = Mn [ D m 1 X (C +Z ) + rd m 1 m 1 X = 1 2 R x m x 1 2 (C + D X, S X = 0, S X 0, = 1,2,3,, m. D buyer s annual demand rate D quantty purchased per year from the th suppler Q order quantty per perod Q order quantty per perod from the th suppler A fxed/order cost for the th suppler X percent of Q assgned to the th suppler S th suppler s setup cost p annual producton rate of the th suppler z varable cost for each product of the th suppler C purchasng prce of each product from the th suppler m number of supplers r annual nventory holdng cost rate T tme of a buyer s perod T tme of consumng an ordered quantty of the th suppler D, A, S, p, z, c, q, and r are known. NUMERICAL EXAMPLE Suppose a purchase manager would lke to buy one product from three supplers wth nformaton as gven n Table1. Demand rate (D) = 1000000 and r = 0.25. Table 1. Suppler s Informaton for stuaton 1 Prce (C ) Fxed Cost (A ) Producton rate (P) Setup Cost (S ) Producton varable cost (Z ) Sup. 1 112 7450 510000 9800 93 Sup. 2 118 6120 670000 9400 89 Sup. 3 114 6590 450000 8600 90 Table 2. Values of varables calculated by LINGO (by Ghodyspour & O Bren) Cases X 1 X 2 X 3 Q Buyer s Annual cost (BAC) Suppler s Annual cost (SSAC) Annual supply chan total cost (ASCT) 1 0.51 0.04 0.45 52248 113868109 92597737 206465846 2 0.51 0.49 0 43778 115564425 91912656 207477081 5 0 0.55 0.45 41972 116811210 90302159 207113369 For stuaton 1, we used hdden and tral method, and regulate the values of Prce (C ) and Product varable cost (Z ). And found that an optmum stuaton s acheved when ncreasng the values of Z by 2 and decreasng C by 3. Z ) P copyrght FACULTY of ENGINEERING HUNEDOARA, ROMANIA 79
Table 3. Values of varables calculated by LINGO for stuaton 1 93,89,90 95,91,92 97,93,94 99,95,96 101,97,98 112,118,114 109,115,111 106,112,108 103,109,105 100,106,102 BAC 113868100 110859000 107849900 104840800 101831700 SSAC 92597740 94610310 96622880 98635450 100648000 ASCT 206465840 205469310 204472780 203476250 202479700 BAC SSAC 21270360 16248690 11227020 6205350 1183700 Fgure 1. Comparson of values for BAC & SSAC For a centralzed model, the company, whch owns both the buyer and vendor, t s possble for the company to accept the stuaton but n case of decentralzed model the buyer and the supplers are components of dfferent companes and n order to optmze the whole supply chan, one of the above case should be accepted. The graph shows a pont where the value of BAC s equal to SSAC but pont to be noted that C can t be more than Z so there should be mutual understandng between the suppler and buyer. The second part of the paper s to nclude the out of stock and excess nventory costs, so that the soluton of the problem could be more practcal. The out of stock costs nclude the beneft of the number of products that has to be selled and excess nventory costs nclude annual nventory holdng costs whch has to carry for the next year. As t s shown n Table 2 that the supply chan total cost for case 2 as compare to case 1 has been reduced whch s contrary to the results for Ghodyspour & O Bren. Table 4. Values of Varables calculated by LINGO after addton of out of stock and nventory costs Annual Supply Chan total cost after addton of Out of Stock and Excess Inventory Costs ASCT AIHC ACST+AIHC Out of ACST+OOS Total Stock(OOS) Case 1 206465846 17819550 224285396 224285396 Case2 207477081 5172300 212649381 212649381 Case 3(Excess Inventory) 207113369 2354000000 2561113369 2561113369 SUPPLY DEMAND MODEL What prce for a commodty results n equal supply and demand? We wll solve by graphng. Suppose the supply curve for a commodty s x=5p 150, where p s the per unt prce n Euro and x s the number of unts of the tem that wll be suppled to the market prce. For example, at a prce of 50(p=50) the supply wll be x=5(50) 150=100 unts, whle at p=100 the supply wll be x=5(100) 150=350 unts. Further, suppose the demand curve for ths commodty s x=375 2,5p, where agan p s the per unt prce and x represents the number of unts of the tem that wll be sold at the prce. For nstance, at a prce of 50, the demand wll be x=375 2,5(50)=250, whle only x = 375 2,5(100)=125 unts wll be demanded at a prce of 100. These calculatons show that demand exceeds supply at a prce of 50 whereas supply exceeds demand at a prce of 100. In each case, then, pressure for a change n prce s created. A stable occurs at a prce for whch supply and demand are equal. We therefore ask f there s a per unt prce p 0 for whch supply equals demand. Such a prce, f t exsts, s called an equlbrum prce. 80 Tome X (Year 2012). Fasccule 2. ISSN 1584 2665
Now suppose there s an equlbrum prce p 0 and that x 0 s the correspondng supply. Then (p 0, x 0 ) must le on the supply curve, that s, x 0 =5p 0 150, or 5p 0 x 0 =150. But snce p 0 s an equlbrum prce, supply equals demand, and so (p 0, x 0 ) must also le on the demand curve, that s, x 0 =375 2,5p 0, or 2,5p 0 +x 0 =375. We conclude that f p 0 s an equlbrum prce and x 0 s the common number of unts suppled and demanded at that prce, then (p 0, x 0 ) s a soluton to the system of lnear equatons 5p x=150 [supply] 2,5p+x=375 [demand] The graph of ths system n the P x plane appears n Fgure 2. The two lnes have a unque pont n common whch we estmate to be (70,200). Thus, the system has a unque soluton, and the equlbrum prce s about 70 Euro per unt. CONCLUSION & FUTURE WORK Fgure 2. Geometrc soluton to the system The major am of the work presented n ths paper s to provde a cost effectve approach that would enable manufacturng organzatons to gan compettve edge n the global market by coordnatng between suppler and buyer to create a wn wn stuaton for a decentralzed model. Integratng and coordnatng the operatons of such organzatons for mproved awareness and responsveness to the changes n the manufacturng envronment. To make the problem more practcal and to acheve ths, out of stock & excess nventory costs have been added. Suppler selecton s a multple crtera decson makng problem that ncludes both qualtatve and quanttatve crtera. These tangble and ntangble factors are not equally mportant. In real cases, many nput data are not known precsely for decson makng. Further extensons to ths work could be carred out to revew the overall effect of dfferng capacty constrants throughout the supply chan. In addton, further research should be carred out to nvestgate the varyng cost of capacty n these systems, whether t s overtme costs or addtonal resource costs. Further extensons to ths paper should ental a revew on the effect of dfferng demand patterns on the overall supply chan performances. Ths s n contrast to the fxed demand patterns used n ths study. The research only nvolves one product, further extenson can be made by ntroducng multple products, n such case products are substtuted or the capacty s lmted. A dscounted cost can also be ntroduced based on the orderng quantty from the supplers. Other numercal softwares can also be used, for example WnQSB whch can also solve smlar problems (stock theory, producton plannng, decson problems, Inventory Theory and System, Job Schedulng) and the results can be compared wth the one whch we obtaned after solvng wth LINGO. REFERENCES [1.] Banerjee, A., 1986. A jont economc lot sze model for purchaser and vendor. Decson Scences 17, 292 311. [2.] Ben Daya, M., Al Nassar, A., 2008. An ntegrated nventory producton system n a three layer supply chan. Producton Plannng and Control 19 (2), 97 104. [3.] Chrstopher, M., Lee, H., 2004. Mtgatng supply chan rsk through mproved confdence. Internatonal Journal of Physcal Dstrbuton and Logstcs Management 34 (5), 388 396. [4.] Ganeshan, R., 1999. Managng supply chan nventores: a multple retaler, one warehouse, multple suppler model. Internatonal Journal of Producton Economcs 59, 341 354. [5.] Ghodsypour, S.H., O Bren, C.O., 1998. A decson support system for suppler selecton usng an ntegrated analytc herarchy process and lnear programmng. Internatonal Journal of Producton Economcs 56 57 (1 3), 199 212. [6.] Ghodsypour, S.H., O Bren, C.O., 2001. The total cost of logstcs n suppler selecton, under condtons of multple sourcng, multple crtera and capacty constrant. Internatonal Journal of Producton Economcs 73 (1), 15 27. [7.] Ghodsypour, S.H., O Bren, C.O., 2007. Optmzng whole supply chan beneft versus buyer s beneft through suppler selecton. Internatonal Journal of Producton Economcs 121 (2009), 482 493. [8.] Goyal, S.K., 1977. An ntegrated nventory model for a sngle suppler sngle customer problem. Internatonal Journal of Producton Research 15, 107 111. [9.] Goyal, S.K., 1988. A jont economc lot sze model for purchaser and vendor: a comment. Decson Scences 19, 236 241. [10.] Goyal, S.K., 1995. A one vendor mult buyer ntegrated nventory model: a com ment. European Journal of Operatonal Research 82, 209 210. [11.] Graves, S. C., &Wllems, S. P. (2005). Optmzng the supply chan confguraton for new product. Management Scence, 51(8), 1165 1180. copyrght FACULTY of ENGINEERING HUNEDOARA, ROMANIA 81
[12.] Goyal, S.K., Nebebe, F., 2000. Determnaton of economc producton shpment polcy for a sngle vendor sngle buyer system. European Journal of Operatonal Research 121, 175 178. [13.] Hll, R.M., 1997. The sngle vendor sngle buyer ntegrated producton nventory model wth a generalzed polcy. European Journal of Operatonal Research 97, 493 499. [14.] Hll, R.M., 1999. The optmal producton and shpment polcy for the sngle vendor sngle buyer ntegrated producton nventory problem. Internatonal Journal of Producton Research 37, 2463 2475. [15.] Khouja, M., 2003. Optmzng nventory decsons n a mult stage mult customer supply chan. Transportaton Research part E 39, 193 208. [16.] Km, T., Hong, Y., Lee, J., 2005. Jont economc producton allocaton and orderng polces n a supply chan consstng of multple plants and a sngle retaler. Internatonal Journal of Producton Research 43 (17), 3619 3632. [17.] Manthou, V., Vlachopoulou, M., Folnas, D., 2004. Vrtual e chan (VeC) model for supply chan collaboraton. Internatonal Journal of Producton Economcs 87 (3), 241 250. [18.] Nurmlaakso, J., 2008. Adopton of e busness functons and mgraton from EDI based to XML based e busness frameworks n supply chan ntegraton. Internatonal Journal of Producton Economcs 113 (2), 721 733. [19.] Tagaras, G., & Vlachos, D. (2001). A perodc revew nventory system wth emergency replenshments. Management Scence, 47(3), 415 429. [20.] Park, S.S., Km, T., Hong, Y., 2006. Producton allocaton and shpment polces n a multple manufacturer sngle retaler supply chan. Internatonal Journal of Systems Scence 37 (3), 163 171. [21.] Van Der Veen, J.A.A., Venugopal, V., 2005. Usng revenue sharng to create wn wn n the vdeo rental supply chan. Journal of the Operatonal Research Socety 56 (7), 757 762. [22.] Yung, K.L., Tang, J., Ip, A.W.H., Wang, D., 2006. Heurstcs for jont decsons n producton, transportaton, and order quantty. Transportaton Scence 40 (1), 99 116. ANNALS OF FACULTY ENGINEERING HUNEDOARA INTERNATIONAL JOURNAL OF ENGINEERING copyrght UNIVERSITY POLITEHNICA TIMISOARA, FACULTY OF ENGINEERING HUNEDOARA, 5, REVOLUTIEI, 331128, HUNEDOARA, ROMANIA http://annals.fh.upt.ro 82 Tome X (Year 2012). Fasccule 2. ISSN 1584 2665