Sunteți pe pagina 1din 5

Study on the Reliability of Hot Standby Repairable Supply System Based on Markov Model

Shuhang Ren, Cunlu Zhang


AbstractStarting from the analysis of cooperation relations of members in the supply system, the paper studies the reliability of repairable supply system. Based on Markov model, a reliability evaluation method of repairable supply system is given. The transient and steady reliabilities of hot standby repairable supply system with a main supplier and a spare supplier are found. The conclusion is tested by a case study. The contribution is that it provides a quantitative method to increase the understanding of dynamic reliability of enterprise supply system.

Index Terms Hot standby supply system, Markov model, Repairable, Reliability
I.

n 2000, Ericsson withdrew from mobile phone market because its single supplier of one of the core components suffered from fire disaster, which in the meantime aroused public concern about the reliability of supply system. Nowadays, as the commercial competition is becoming fiercer and fiercer, the supply system management which has been playing a more and more important role in business activities has become an indispensible part in competitive advantages. Therefore, its getting increasingly urgent to evaluate and improve the reliability of enterprise supply system. To study the reliability of supply system is to evaluate the general effect after taking measures. Thus, the standard of reliability evaluation, i.e. the goal of reliability control, which is based on the planned target of supply system, should be set first. Mr. Zhang Cunlu has offered the principles to set the standard of the reliability of supply system, including the available operation target, planned target, the expected customer service level, the present management level, etc [1]. However, how to set the standard of reliability evaluation is not the key point in this paper. A number of domestic and overseas scholars have studied on the supply chain and the reliability of its operation activities from many perspectives and have obtained many achievements. Document 2 has offered an evaluation matrix of the reliability of supply chain, regarding various factors affecting reliability,
Shuhang Ren, is with management science department, school of management,XiamenUniversity,CO361005 China(mobile phone:13950159344,e-mail:renshuhang2006@yahoo.com.cn). Cunlu Zhang, is with management science department, school of management, Xiamen University,CO361005 china. He is the viceprofessor.(mobile phone:13850002630,e-mail:zclu@sina.com).

INTRODUCTION

supply chain. Then, the reliability level of the whole supply chain can be worked out according to general series principle. Document 3 simply describes the single-level supply chain and evaluates the reliability of multi-level supply chain based on the Markov process. Furthermore, some suggestions on how to avoid risk in supply chain and improve reliability are also provided. Document 4 puts forward the general structure of suppliers network risk management as well as the solutions under complex network environment and demonstrates the challenge of suppliers network cooperation to the supply chain risk management. Document 5 starts from the workflow of distribution service and applies relevant theories of Markov process to put forward a forecast method of reliability of distribution service based on Vector Markov Chain. Thus, distribution service under steady state and the reliability of its business processes are obtained. Document 6 first derives the expression of reliability level of single-part repairable system based on Markov process, and deduces a formula of availability and reliability of the standby redundancy repairable system. According to the available document research, studies on the reliability of hot standby repairable supply chain are quite few at present. Besides, on analyzing and evaluating the reliability of supply chain, the present documents mainly concern two aspects: one is to make comprehensive evaluations in terms of external factors index, including the concrete method of analytical hierarchy process, fuzzy comprehensive assessment and so on, or to make analyses in terms of internal factors, such as fault tree analytical method; the other is to analyze the structure of supply chain according to the theory of system reliability and divide it into serial system, parallel system, compound system, redundant system and network system to study system reliabilities on the basis of their internal logical relationships. However, the present researches show that the first method is more often used, although its subjectivity of weight setting to some extent makes inaccurate results; while the second one seldom differentiates the repairability of supply chain system, which means it doesnt consider the reoperation after the systems repair. In fact, supply chain is not able to rebuild its system immediately encountering malfunction or invalidation, so its more practical to resort to scientific repairing, which has just offered a cut-in point for my paper to evaluate the reliability of supply system under repairable circumstances. Therefore, starting from the corporative relations of members of the supply system, this paper examines the

978-1-4244-3662-0/09/$25.00 2009 IEEE

318

reliability of hot standby repairable supply system which consists of a main supplier and a spare supplier. Besides, based on the Markov model which emphasizes state conversion instead of internal mechanism, the paper also brings forward a calculation method of evaluating reliability of hot standby repairable supply system. This method not only provides some evidence for evaluating the present reliability level, but also compares the result with the previous evaluation standard or much higher target to make sure whether the deviation has been controlled in a certain scope. II. THE DEFINITION OF RELIABILITY OF SUPPLY SYSTEM From the perspective of system engineering, reliability refers to the probability that parts, components, products or systems complete their prescribed functions smoothly under the set condition within the fixed time [7]. Although many experts and scholars have studied on the reliability of supply chain and provided its definitions from various points of view, there hasnt been an authoritative definition so far. Thomas (2002) claims that supply chain reliability refers to the probability of delivering supply to the key points of supply chain system to satisfy demands [8]. Mu Dong and Du Zhiping (2004) defines the term as the ability of supply chain to complete the demand of orders under the set time and condition, which is generally presented as the accessibility and the minimum acceptable service level of supply chain [9]. Based on the theory of system reliability engineering, Mr. Liu Yuanhong (2005) defines it as the measurement of supply chain systems fault-free working ability, which refers to the normal operation ability within fixed time on the basis of completely competitive market [10]. According to the trading theory and system reliability theory, Tian Guiliang (2007) defines supply chain coordination reliability as the ability of each enterprise in the supply chain to fulfill their order demands to ensure the smooth operation of the whole supply chain, on the condition that the supply chain system is not disturbed by external factors [11]. On the basis of the above definitions of system engineering and supply chain reliability, this paper adopts the following definition of supply system reliability: it refers to the probability of suppliers to deliver correct materials to customers under the right time and place, with right product package as well as right document files. III. STUDY ON THE RELIABILITY OF HOT STANDBY REPAIRABLE SUPPLY SYSTEM BASED ON MARKOV MODEL A. Markov Model and the Basic Hypotheses Markov model, exploring the interconversion among various states, was put forward by a Russian named Markov in 1907. If at a certain moment, the conversion probability from one state to another is only related to the present state instead of the previous state, that is it has nothing to do with the finite previous state, then this process is called Markov process [7].

For the requirement of model and calculation, the paper provides the following hypotheses: Hypothesis one: This paper assumes that supply system consists of a main supplier and a spare supplier. As manufacturers core components will involve various factors such as suppliers patent, property right and high-level technology, single supplier strategy should be adopted to maintain the partnership or strategic alliance with suppliers. To manufacturers, the products and services that strategic suppliers provide are of great importance and value and will probably have impact on their own products and operation processes or even influence the ultimate capacity to satisfy the demands of custo mers. Because of their str ong competitiveness, these products and services which are generally aimed at manufacturers specific requirement have achieved high personalization and differentiation. However, there are comparatively less strategic suppliers who can satisfy their own demands, and their conversion cost is also pretty high. Therefore, in consideration of the risk, spare suppliers should be set to deal with the problems incase main suppliers malfunction. According to the above considerations, it is necessary to evaluate the reliability of supply system which c o n s i s t s o f a ma i n s u p p l i e r a n d a s p a r e s u p p li e r . Hypothesis two: This paper assumes the supply system is repairable. Nowadays, as competition has become globalized, manufacturers and suppliers gradually resort to cooperation strategy. Thus the factors influencing supply system reliability also include the long-term cooperation ability of the chosen supplier. As for those who have encountered problems, manufacturers tend to choose joint effort to improve suppliers operation level and enhance their partnership instead of changing suppliers like before. It is the strategic partnership of upstream and downstream members that makes their hand-inhand improvement, with the goal of long-term cooperation and mutual win. From this point, supply system can be regarded as a repairable system, which is alternated by normal operation and breakdown maintenance. Hypothesis three: As an enterprise, suppliers will take some risk to cooperate with manufacturers in the supply chain. Suppliers may be at standby state in one supply chain, but at working state in another, so this paper assumes that suppliers are possible to malfunction no matter what kind of state they are at in supply chain. That is to say, what we are discussing here is hot standby repairable system. We define the main supplier in supply system as unit 1 and spare supplier as unit 2, with a constant failure and improvement rate. At working state, their failure rates are respectively 1 and 2 , and improvement rates 1 and 2 . As the system is hot standby system, failure rate is 3 at standby state( 1>1,2 ,3 >0 1> 1,2 >0 . Hypothesis four: If unit 1 (the main supplier) fails and unit 2 (the spare supplier) alternates to function, manufacturers and

978-1-4244-3662-0/09/$25.00 2009 IEEE

319

suppliers cooperate to solve problems and satisfy demands, then this supply system is standby repairable system. Hypothesis five: As long as one unit is working, the supply system is at normal state. However, if both unit 1 and unit 2 are ineffective at the same time, then the supply system malfunctions. Hypothesis six: If unit 2 fails and unit 1 is still at working state, manufacturers should contact unit2 immediately and solve problems together. Hypothesis seven: The successful maintenance of unit 1 and failure of unit 2 will not happen simultaneously and the two will not fail at the same time as well. B. Evaluation on the Reliability of Hot Standby Repairable Supply System From the above hypotheses, we can figure out four possible states of supply system: 0: unit 1 is at working state while unit 2 at standby state. Now the system is normal; 1: unit 2 is at working state while unit 1 at the state of being improved. Now the system is normal; 2: unit 1 is at working state while unit 2 at the state of being improved. Now the system is normal; 3: unit 1 is at the state of being improved and unit 2 is the same. Now the system is ineffective. The result of supply system must be one of the four kinds if states upwards at a certain point. However, the result of supply system may transfer from one state to another during a period of time. Besides the result of every time the state of supply system is only related with the process of current operation of supply system, and will not influence the other operation of supply system. Therefore the state transition of supply system can be regarded as Markov Chain. Figure 1 shows the model of state transition within the time of t :

1(1+3)t 1t 3t 0 0 2t 1t 1(2+1)t P(t ) = 2t 0 1( +2)t 1t (1) 1 2t 1t 1(2+1)t 0


( Pi j ( t ) i , j=0,1,2,3 indicates the transition probability from state i to j ) On the basis of micro-modulus transfer matrix, we get the following state transition equation from t 0 :
(1+3) 1 3 0 0 1 (2 +1) 2 P (t ),P (t ),P (t ),P (t ) 1 2 3 0 0 (1+2 ) 1 2 (2) 2 1 ( 2 +1) 0 = P (t ),P (t ),P (t ),P (t ) A = P' (t ),P' (t ),P' (t ),P' (t ) 1 2 3 1 2 3 0 0 Pj (t ) , j=0,1,2,3 shows the supply system transient

probability at any moment t, i.e. P (t ) A = P' (t ) (3) The reliability of supply network can be analyzed from the following two aspects. If the systems initial state is given, the transient probability of any normal working state can be figured out from the following procedure [12]: 1) After making Laplace transform to equation (3), we get P ( s )( sI A )= P (0) ( P (0) means the initial state of system), that i s P ( s ) =P (0)( sI A )1 (4) 2) Work out ( sI-A ) and its inverse matrix ( sI-A )1 3) Expand ( sI-A )1 through partial fraction method, and insert it in formula (4). After making Laplace transform, well finally get the result of P (t )=[ P0 (t ), P (t ), P2 (t ), P3 (t )] . Thus, the 1 transient reliability should be (5) Although there is a solution to work out transient reliability, this method involves matrix inversion which is generally complicated, so its not easy to derive the theoretical model solution of transient reliability. However, as for some special data, matrix inversion can be easily solved through EXCEL and MATLAB, and transient reliability is then figured out based on the above calculation processes. Below is the solving procedure of steady reliability when the system is at a steady state. The reliability of supply system under steady state refers to the probability of supplied state when the operation of supply system reaches a steady state, which means no matter what state the initial supply system is, its probability of being various states tends to be steady after a long-time operation. As a constant, this probability has nothing to do with the initial state. Let lim Pj (t ) = j , j = 0,1, 2, 3
t
R (t )= P0 (t )+ P (t )+ P2 (t ) . 1

1 ( 1 t + 3 t ) 0

1 ( 2 t + 1 t )

1 t
1t
2t

3 t
2

2t

2 t
3

1 t
1 t

1 (1t + 2 t )

1 (1t + 2 t )

Figure1The Model of State Transition If the malfunction is repaired immediately, then a micromodulus transfer matrix within the time of t ~ t + t is derived from figure1:

978-1-4244-3662-0/09/$25.00 2009 IEEE

320

as lim Pj (t ) = 0, j = 0,1, 2, 3 , t (6) Let t , then the system tends to be steady, and the above equation can be transformed into the following linear equation:
0 +1+ 2 + 3 =1

to (8) we can get the varies of the probability of supply chain system at the state of 0 and the steady reliability ( R ( ) ), While 1 2 and 3 vary (see table 2). TABLE 2. The Varies of The Probability of Supply Chain System at the State of 0 and the Steady Reliability R ( )
1 0
R ()

3 0 ( +3) 1 1 ( + ) 0 2 1 2 1 =[0,0,0,0] [0,1,2,3] 2 0 (1+2) 1 2 1 (2+1) 0


Obviously, the four equations in the system are linearly dependent, so we have to add 0 +1+ 2 + 3 =1 and solve the linear system as follows:
1 2 ( 1 + 2 + 1 + 2 ) 0 = ( 1 + 1 ) [ ( 2 + 2 ) ( 1 + 3 + 2 ) + 1 ( 3 + 2 ) ] 1 2 ( 1 + 3 + 1 + 2 ) 1= ( 1 + 1 ) [ ( 2 + 2 ) ( 1 + 3 + 2 ) + 1 ( 3 + 2 ) ] 1 ( 1 2 + 2 3 + 1 3 + 2 3 ) 2= ( 1 + 1 ) [ ( 2 + 2 ) ( 1 + 3 + 2 ) + 1 ( 3 + 2 ) ] 1 3 ( 2 + 1 ) + 1 2 ( 1 + 2 ) 3 = ( 1 + 1 ) [ ( 2 + 2 ) ( 1 + 3 + 2 ) + 1 ( 3 + 2 ) ]

(7)

0.33 0.3 0.27 0.24


2

0.462736 0.475337 0.488633 0.502685


0

0.894086 0.901204 0.908702 0.916613


R ()

0.47 0.44 0.41

0.462736 0.464374 0.466065 0.467814


0

0.894086 0.897005 0.900022 0.90314


R ()

(8)

0.38
3

0.4 0.37 0.34 0.31

0.462736 0.473282 0.484319 0.495884

0.894086 0.895783 0.89756 0.899421

Thus, the steady reliability of supply system should be: (9) Compare the calculated steady reliability with the previously set reliability evaluation standard to check the reliability level of repairable supply system at a certain time. (see table 1)
R ( ) = P0 ( ) + P ( ) + P2 ( ) = 0 + 1 + 2 1

From the table 2,wo can get that the steady reliability ( R ( ) ) has small changes , and the 1 3 can cause the 0 and R ( ) varying more than that of the 2 ,while 1 2 and 3 decline. From the results of above, we can find that if the enterprise wants to improve the reliability of its supply system, they need to chose the supplier with high reliability, and improve the spare suppliers reliability at standby state by putting emphasis on the spare supplier relationship management. V. CONCLUSION Through quantitative analysis and a case study, the conclusion is that the method of reliability evaluation of hot standby repairable supply system, which is based on Markov model, can not only evaluate the reliability of supply system correctly when the supply system operates steadily, but also analyze the influence of the reliability of supply system according to the states of supply system and find out the key point of improve the reliability of supply system. Furthermore, this study offers theoretical and empirical approach to the enterprise that want to evaluate and improve reliability of supply system. In view of the complexity of reliability of repairable supply system, this paper only studies the upstream link of the systemthe reliability of supply chain system, but the reliability of large-scale supply chain needs further study.

TABLE1. the Standard of Reliability Evaluation 0.1<R<0.3 0.3<R<0.8 R>0.8 Reliability R<0.1 (R) Reliability The lowest Low level Medium High level level(highest (high risk) level level risk) (medium (low risk) risk) IV. CASE STUDY There is a supply system A of a enterprise. According to the historical data of operation, the parameters are estimated. We can get that: 1 is 0.33, 1 is 0.83, 2 is 0.47, 3 is 0.47,and 2 is 0.75.Accoding to (8),we can get

0 = 0 .4 6 1 = 0 .1 8 2 = 0 .2 5 3 = 0 .1 1 So the reliability of hot standby repairable supply system is R()=0+1+2=89%,and we can get the level of reliability is high. The probability of unreliability of supply system may only caused by the spare supplier is 0.36, the probability of unreliability of supply system caused by the spare supplier and the main supplier is 0.11. The state of 0 shows that unit 1(main supplier) is at working state while unit 2(spare supplier) at standby state. And it also shows that the system is under the ideal state. According

978-1-4244-3662-0/09/$25.00 2009 IEEE

321

REFERENCES
[1] C. Zhang. Supply chain risk management. Beijing: Tsinghua University Press,2007. [2] W. Ding, K. Liu, G. He. Study on Risk of Supply Chain. China Safety Science Journal, vol.13,no.4,pp.64-66,2003. [3] J. Wang, W. Zhang. Analysis on Reliability of the Supply Chain. China Safety Science Journal, 13(11): 7376,2003. [4] J. Hallikas, I. Karvonenb, U. Pulkkinenb, M. Virolainen. Risk management processes in supplier networks. Production Economics,no.90,pp.47-58,2004. [5] X. Zhang, J. Wu. Forecast of reliability of distribution service based on vector Markov chain. Journal Of systems Engineering,vol.22,no.3,pp.:300-304, 2007.

[6] Z. Wu, C. Guo, H. Zhao. Reliability-modeling for the repairable system based on Markov process. Journal of Dalian Maritime University, vol.33,no.1,pp.13:-16, 2007. [7] Y. Guo. Principle of reliability engineering. Beijing: Tsinghua University Press,2002. [8] M. U. Thomas. Supply chain reliability for contingency operations. Proceedings Annual Reliability and Maintainability Symposium, no.1,pp.61-67,2002. [9] D. Mu, Z. Du. On Reliability of Inherency and Operation in SC . Logistics Technology,no.12,pp.72-74, 2004. [10] Y. Liu, M. Luo, Z. Liu. Management on Reliability of the Supply Chain. Modern management Science,no.5,pp.15-16, 2005. [11] G. Tan, C. Xu. Study On Reliability management of Supply Chain Coordination. Productivity Research,no.23,pp.109-112, 2007. [12] B. Song. Reliability Design and Analysis of System. Xi an Northwestern Polytechnical University Press, 2000.

978-1-4244-3662-0/09/$25.00 2009 IEEE

322

S-ar putea să vă placă și