• [email protected]
  • +971 507 888 742
Submit Manuscript
SciAlert
  • Home
  • Journals
  • Information
    • For Authors
    • For Referees
    • For Librarian
    • For Societies
  • Contact
  1. Information Technology Journal
  2. Vol 12 (4), 2013
  3. 688-695
  • Issues
    Online First Current Issue All Issues
  • Information About
    Aims and Scope Editorial Board Guide to Authors Article Processing Charges
    Submit a Manuscript

Information Technology Journal

Year: 2013 | Volume: 12 | Issue: 4 | Page No.: 688-695
DOI: 10.3923/itj.2013.688.695
crossmark

Facebook Twitter Reddit Linkedin E-mail
Research Article

Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts

Lei Sun
College of Civil Aviation, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, People Republic of China

Hongfu Zuo
College of Civil Aviation, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, People Republic of China

ABSTRACT


Owing to the expensive and sporadic demand nature of aircraft spare parts, the stock level of spare parts was difficult to confirm. How to maximize the availability of the aircraft fleet with cost constraint for spare parts was an important issue for airlines. A multi-ECHELON inventory model based on METRIC (Multi-ECHELON technology for recoverable items control) was proposed by allocating spare parts among the bases of airlines in this study. The model based on system approach instead of item approach to determine aircraft spare parts stock level in a multi-ECHELON system. First, the key system measure was selected, then, the Negative binomial distribution was employed to more accurately reflect the variance in part failure processes when the value of variance-to-mean ratio was bigger than one, Next, the marginal analysis method was applied to find the optimal position of spare parts among bases and depot, finally, two examples were given and the first one showed that the negative binomial distribution was more reasonable than Poisson distribution to describe the non-stationary demand process. The second one showed that the models in this paper were engineering applicability for airlines and could provide an effective theoretical and technical support.
PDF Abstract XML References Citation

Keywords


  • inventory optimization
  • system approach
  • multi-ECHELON
  • Spare parts
  • negative binomial distribution

Article History

Received: September 06, 2012;   Accepted: December 21, 2012;   Published: January 22, 2013

How to cite this article

Lei Sun and Hongfu Zuo, 2013. Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts. Information Technology Journal, 12: 688-695.

DOI: 10.3923/itj.2013.688.695

URL: https://scialert.net/abstract/?doi=itj.2013.688.695

INTRODUCTION


Equipment-intensive industries such as airlines, nuclear power plants, played an ever more important role in society. The availability of such systems may strongly affect daily operations. As a consequence, it has required large quantities of spare parts to guarantee their availability which in turn resulted in excessive cost. The aviation industry, for example, it must carry about hundreds billion dollar each year to stock the spare parts they need to keep their airplanes flying (Rambau and Schade, 2010). So it was very important for airlines to determine the stock level with reasonable inventory investment.

The importance of service parts management had increased in the past decades. Many models for these kinds of stock allocation problem had been developed. They could be classified into two streams. The works of Yanagi et al. (1997), Wong et al. (2002) and Van Harten and Sleptchenko (2003) belonged to the first stream. They modeled the problem as a multi-dimensional Markovian problem. All research in the second stream was based on the well-know METRIC model proposed by Sherbrooke (1968) for the US Air force, including the works by Slay (1984), Graves (1985), Sherbrooke (1986, 2004), Axsater (1990), Zamperini and Freimer (2005) and Wong et al. (2006). These models were largely focused on spare parts inventories and was considered as system approaches that aiming at a high availability of complete technical systems, as opposed to more classical inventory management approaches, that were primarily directed towards a high availability of individual items. Compared to the models in the second stream, the models in the first stream give more exact results than the METRIC type models. However, they were more difficult to solve because of the huge multi-dimensional state spaces involved same as the multi-objective optimization applying evolutionary algorithms. Due to the complex nature of aircrafts’ repair process, the METRIC model which could be evaluated and optimized within reasonable computation times, was chosen to solve this problem.

On the early version of the METRIC model, the number of LRUs in the pipeline which denoted the number of LRUs in repair or being resupplied from a higher ECHELON, was assumed to follow Poisson distribution (Jiang-sheng et al., 2008; Wang and Kang, 2009; Rui et al., 2011), but Sherbrooke (2004) proved that the value of variance-to-mean ratio of backorder at the depot was bigger than one which meant the Poisson distribution was not suitable for the distribution of the number of LRUs in the pipeline of bases which was composed by the backorders. In this study, the negative binomial distribution was employed to more accurately reflect the distribution of the number of LRUs in the pipeline of bases and a example was given to show that by taking into account not only the mean value but also the variance, it could obtain a much better estimation for the non-stationary demand.

The aim of this paper was to determine the stock level of those expensive but low-usage aircraft spare parts, LRUs (Line Replaceable Units), in multi-ECHELON maintenance organization with minimum cost. This paper analyzed LRUs demand generation, transfer process and the fault repair process in multi-ECHELON maintenance support organizations. Based on the research conclusion, the spares stock level was given in a numerical study.

MODEL DESCRIPTION AND ASSUMPTIONS

An airline company usually kept a central inventory of parts in its depot. Additionally, it also kept smaller ‘outstation’ inventory at the other airports bases where its aircraft had regularly scheduled landings and departures.

Generally, after an aircraft had landed at a base, engineering inspection was carried out prior to the next departure. When a LRU on an aircraft become defective, it should be removed and replaced by a serviceable one from the local stock (if available). If there was no available one in the local inventory for the replacement, a backorder was established and the aircraft will remain on the ground and will be delayed until an incoming flight brings a replacement part from the depot. The failed part could be repaired at the base for some minor problems but for the more serious problems it should be sent to the depot or to a special repair facility. At the same time, a functioning LRU was sent from the depot to the base. But if the depot was empty, a depot backorder was established. This does not necessarily imply that an aircraft was grounded but the risk of backorders at the base increases.

The transportation time for a defect LRU from a base to the central depot was assumed to be deterministic and known and the same was assumed for the transportation time from the central depot to a base. For simplicity, it was assumed that there was no difference between the bases in this respect.

The repair time for a defect LRU at the depot was assumed to be a random variable with expected value. An important assumption (approximation) in the model was that these repair time were independent and equally distributed. According to Palm’s theorem, if demand for an item was a Poisson process with annual mean m and the repair time for each failed unit was independently and identically distributed according to any distribution with mean T years, then the steady-state probability distribution for the number of LRUs in repair had a Poisson distribution with mean mT.

There were the following variable definitions for the models:

τij = Average order and ship time from the depot to line base j for LRUi
p(x) = Probability that the demand for a given LRU during a predefined time
I = Index of LRU type i, i = 1, 2,..., I, where, I is the total number of LRUs in the system
λij = Average annual demand of the LRUi at base j
j = Index of base j, j = 0, 1,..., J, where, 0 is the depot and J is the total number of bases
vij = Average repair time (in one year) of the LRUi at base j
ρij = Probability of repair of the LRUi at base j
μij = Average pipeline at base j, represents the average demand for the LRUi under repair or resupply

MODEL TECHNIQUE

In this section we briefly discussed the models of multi-ECHELON inventory for airlines based on the METRIC theory. The problem will be solved in three steps:

Step 1: The definition of the expected backorder (EBO) was given and the relationship between the fleet availability and EBO was deduced
Step 2: For a single LRU, inventory models were developed to allocate stock level between the depot and the bases and marginal analysis was used to find the optimal allocation of stock levels between the bases and the depot with the cost constraints
Step 3: All LRUs were combined on a system using marginal analysis. It was shown how to construct an optimal cost-backorder curve for all LRUs as a system

Mean and variance for the backorders: At a given randomly chosen time, there was a famous balance equation for any spare which must be at one of the state. The balance equation was the basis for all of the analysis to come:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(1)

The stock level, s, was the stock level; DI was the number of LRUs coming from repair and resupply, OH was the number of LRUs currently available in the inventory (on hand), BO was the number of backorders. They were all natural random variables which could only take on non-Negative integer values, Moreover, at each time at least one of BO and OH was zero. One of the variables changed, the others would change.

Therefore, BO and OH could be expressed as the following functions of X and s:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(2)

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(3)

Suppose p(x) was the distribution of the number of LRUs coming from repair and resupply. The expected number of backorders, EBO(s), was thus:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(4)

And the recursive expression of Eq. 4 was:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(5)

Referring to the definition of variance, the variance of backorders (VBO) was get:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(6)

The deriving expression of E[BO2(s)] is:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(7)

So the recursive expression of Eq. 6 is:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(8)

Objective function and constraints: Availability, A, was the expected percent of the aircraft fleet that was not down for any spare. Treating failures of LRUi in different installed locations as independent, the probability that no randomly selected element was missing a LRU is:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(9)

N was the number of aircrafts; Zi was the number of the LRUi that installed on an aircraft; i was the number of LRU type, An aircraft will be available only if there was no backorders for any of the occurrences of LRUs, An aircraft will be available only if there was no backorders for any of the occurrences of the ith LRU (which accounts for the exponent Zi) and for any of the LRUs, Taking logarithms in Eq. 9:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(10)

So the relationship between fleet availability and EBO(si) was: max (A)_min (sum (EBO(si))) and now the optimal mathematical statement for the aircrafts fleet at all bases was:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(11)

C is the total system cost targets, ci was the cost of LRUi; sij was the stock level of LRUi at base j.

The pipeline and backorders at depot: For a single LRU, the pipeline at a base included the LRUs in repair and resupply and the resupply time depended on the EBO at depot.

The defected LRUs arrive to the bases were assumed to according to a Poisson process, then the average demand at depot was the sum of the LRUs that couldn’t be repaired at the bases with intensity:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(12)

As the repair time had been assumed independent, it follows from Palm’s theorem that the average number of LRUi in the pipeline at depot, Xi0, was a Poisson random variable with:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(13)

The EBO(si0) at depot, with the same recursive equation as Eq. 5 is:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(14)

The variance of backorders was get according to Eq. 6:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(15)

The pipeline and backorders at bases: The pipeline of LRUi at base j consisted of three parts: LRUi in repair at the base j, LRUi on order and LRUi waiting at the depot for backorders. The fraction of all demands at the depot for LRUi that was resupplied to base j was as below:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(16)

So the pipeline could be expressed as:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(17)

When s = 0, EBO(si0) was reduced to the mean of probability distribution and VBO(si0) was reduced to the variance of probability distribution, so the ratio VBO(si0)/EBO(si0) was one. The typical behavior for the ratio was to increase as a function of s to a maximum value slightly larger than s and then decrease asymptotically to one, as shown in Fig. 1. Thus the probability distribution of backorder for LRUs at the depot was not Poisson unless LRUs stock level s = 0. Because μij was composed by backorder function, so Poisson distribution wasn’t suitable unless the stock level at the depot was zero.

The negative binomial was proved that its mean and variance was more suitable for the values of the pipeline than the Poisson by Slay (1984) and later by Graves (1985). So the Negative binomial distribution was used to describe the distribution of the number of LRUs in the pipeline at bases, μij but the function had an added work requiring two parameters rij and pij which could get from the mean and the variance:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(18)

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
Fig. 1: VBO(s)/EBO(s) for various mean values of the Poisson

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(19)

The expression for the variance of the number of LRUs in the pipeline at the base j was got by Sherbrooke (2004):

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(20)

For computational purposes, it was useful to had recursion formulas. The derivation was below:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(21)

The Recursive expression of the negative binomial distribution was below:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(22)

where, 0<p<1 and r cannot be integer.

According to Eq. 4 the expected backorders for LRUi at base j was:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(23)

Calculating and optimization procedure:

Step 1: Start with the depot stock level of zero for one LRU
Step 2: Calculate the number of LRUs in the pipeline for the depot from Eq. 13 and the expected backorders and variance from Eq. 14-15 of the depot
Step 3: Calculate the average number of LRUs in the pipeline for the base j from Eq. 17 and variance from Eq. 20. Then the two parameters r and p could get from Eq. 18 and 19
Step 4: Calculate the expected backorders for each stock level of bases from Eq. 23. Repeat for each base
Step 5: Use marginal analysis to combine the base backorder functions and obtain the minimum backorders for each number of LRU at bases
Step 6: If the level of depot stock was large enough, go to step 7; otherwise, increase the depot stock level by one and go to step 2
Step 7: Find the minimum sum of expected backorders value for each number of the LRU stock
Step 8: Repeats steps 1-7 for each LRU
Step 9: Use marginal analysis to combine the LRU solutions, where the delta of backorder must be divided by the LRU costs

To prove that marginal analysis produces optimal solutions, The EBO function was need to prove convex for any probability distribution:

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(24)

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
(25)

The first difference of EBO was less than or equal to zero and the second difference was greater than or equal to zero which was according with the definition of the convex function.

Since the expected backorder function was convex, the marginal analysis values {EBO(s-1) -EBO(s)}/c were non-increasing. The system backorders were convex also.

NUMERICAL EXAMPLE AND RESULTS

Here, we present two examples. The aim of the first one was to show the reasonable of selecting Negative binomical distribution to describe the non-stationary demand process and the second one elaborated the calculating and optimization process of a multi-ECHELON inventory problem.

Example 1: A real sample of spares demand in 36 months was shown in Table 1. The sample mean was [30x0+3x1+1x(2+3+4)]/36 = 0.33, so the column of (y-mean)2 was the deviation degree from which variance could got 0.778. The column of E/V was the value of variance-to-mean which stated clearly the demand process was non-stationary. Table 2 listed the sample frequency, the probability of Poisson distribution and Negative Binomial distribution adjusted to the sample whose mean was 0.33 and variance was 0.778. From this table could be seen there was 83.3% zero demand which was the low-usage nature of those expensive aircraft spare parts.

Table 1: A real sample of spares demand
Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts

Table 2: Poisson and negative binomial distribution adjust to the sample
Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
Fig. 2: The comparison of Poisson and negative distribution adjust to the sample

Figure 2 was drawn according to Table 2 and it showed clearly that Negative binomial distribution could more accurately reflect the sample whose variance to mean was bigger than one.

Example 2: In this example, there were two organizational levels as described above, two types of LRUs. See Fig. 3.

The input data for LRU1: λ1j = 20 demands/year, v1j = 0.01 years, ρ1j = 0.2, τ1j = 0.01 years, c1 = 5, v10 = 0.025 years; LRU2: λ2j = 10 demands/years, v2j = 0.01 years, ρ2j = 0.1, τ2j = 0.01 years, c2 = 8, v20 = 0.02 years.

Taking LRU1 for example, when the depot stock level, s10 was zero and the following values λ10 = 4x20 H(1-0.2) = 64, μ10 = 64x0.025 = 1.6 were determined from Eq. 12 and 13. It was clearly that EBO(s10= 0) was the mean of Poisson distribution which could be get from Eq. 4 and VBO(s10 = 0) was the variance of Poisson distribution, so they all equaled 1.6. Poisson distribution was still used to calculate the number of LRUs in the pipeline for the bases. Just taking base 1 for example, the number of LRUs in the pipeline was:

μ11 = 20x0.01x0.2+20x0.01x0.8
+[(20x0.8)/64] x1.6 = 0.6

from Eq. 14 which was the expected value of backorder, so EBO(s11 = 0) = 0.6, Then the stock level increased and EBO(s11) was calculated from Eq. 14. The marginal analysis value, EBO(s11-1)-EBO(s11) (ΔEBO(s11)) for any base could be calculated, as shown in Table 3.

The total expected backorders with zero stock was just the sum of the backorders at each base and at the depot as 0.6Δ4+1.6 = 4, seeing the first line of Table 4. Because the cost of one LRU didn’t vary by base, so the delta of EBO didn’t need to divide the first differences by unit cost to apply marginal analysis. When the first spare was allocate to base one, the delta of EBO was 0.6-0.149 = 0.451 and the total EBO was 4.0-0.451 = 3.549, seeing the second line of Table 4. Then the second spare was allocate to base two and the total EBO was 3.549-0.451 = 3.098, seeing the third line of Table 4. Then the third and the forth spare were allocated to base three and base four. The four bases were assumed to be identical, so the stock was allocated to each base successively.

Table 3: The expected backorders of base 1 for the depot stock level 0
Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts

Table 4: Optimal expected backorders between bases for the depot stock level 0
Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts

From the above procedure, the optimal solution was got for the depot stock level zero. Then the depot stock level increased to one and the same procedure need to repeat, when s10 wasn’t zero, the EBO(s10) and VBO(s10) could be got from Eq. 14 and 15. The value of variance-to- mean wasn’t one, so Negative binomial was used to calculate the expected values of backorder of the bases, the two parameters p11 and r11 were calculate from Eq. 18 and 19. The values of these variables for depot stock level 0-6 were shown in Table 5. The expected backorders for the bases, EBO(s1j), calculated by Negative binomial distribution from Eq. 23. When the cost of LRU1 exceeded the total cost, the total stock of LRU1 stopped to increase. Then the following lines of Table 6 could be got. For example, when the total stock level was 10, 5 units were allocated to the depot, 2 units were allocated to base one and the other three bases got 1 unit separately.

LRU2 repeated the same procedure as LRU1 and the optimal allocation for LRU2 could be got. Then using marginal analysis as that used in Table 4 to combine the two units but the delta of backorder must be divided by the LRU cost. Then Table 7 could be got.

The curve of system cost-backorder could be drawn according to Table 7, as shown in Fig. 4. It obviously was convex and the solution was the optimal one. This could be converted into an optimal system availability-cost curve in the same manner also which made the model a very powerful tool for comparing different solutions.

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
Fig. 3: The level of the organization

Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts
Fig. 4: Optimal system cost-backorder curves

Table 5: The values of variables for the depot stock level 0-6
Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts

Table 6: Optimal expected backorders for any stock level
Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts

Table 7: Marginal analysis for each LRU
Image for - Optimal Inventory Modeling of Multi-ECHELON System for Aircraft Spares Parts

For example, when choosing between two components performing the same function, was it more economical from a cost perspective to select the expensive component with low failure rate, or a cheap one with a higher failure rate? Just input their data and see the total backorder result.

DISCUSSION AND CONCLUSION


In this study, it was proved that Negative binomial distribution was more suitable than Poisson for the number of units in the pipeline of bases. The marginal analysis was used to optimize the expensive LRUs inventory with state dependent repair and failure rates. Computational results showed that the proposed approach was efficient in determining the optimal choice of spares for the multi-ECHELON repairable inventory system. The model was not only suitable for confirming initial spare parts list but also well for optimal reallocation and replenishment of existing spares assortments. Further research could be focus on lateral translation, because it could obviously reduce the backorders by pooling the inventory at the same level.

ACKNOWLEDGMENTS


This study was supported by the National Natural Science Foundation and the Civil Aviation Administration of China joint funding key projects (60939003).

REFERENCES


  1. Rambau, J. and K. Schade, 2010. The stochastic guaranteed service model with recourse for multi-echelon warehouse management. Electronic Notes Discrete Mathe., 36: 783-790.
    CrossRefDirect Link

  2. Yanagi, S., K. Hasegawa and T. Yuge, 1997. An approximation to the steady state probabilities of a multi-echelon repair model for a series system. Comput. Ind. Eng., 33: 745-748.
    CrossRef

  3. Wong, H., D. Cattrysse and D. van Oudheusden, 2002. An approximation method for the problem of pooling in a repairable-item inventory system. Proceedings of the 16th ORBEL Conference on Quantitative Methods in Decision Making, (QMD'02), Brussels, Belgium, pp: 75-76.

  4. Van Harten, A. and A. Sleptchenko, 2003. On Markovian multi-class, multi-server queueing. Queue. Syst., 43: 307-328.
    CrossRef

  5. Sherbrooke, C.C., 1968. Metric: A multi-echelon technique for recoverable item control. Operat. Res., 16: 122-141.
    CrossRef

  6. Slay, F.M., 1984. Vary-metric: An approachto modeling multi-echelon resupply when the demand process is poisson with gamma prior. Report AF301-3, Logistic Management Institute, Washington, DC.

  7. Graves, S.C., 1985. A multi-echelon inventory model for a repairable item with one-for-one replenishment. Manage. Sci., 31: 1247-1256.
    Direct Link

  8. Sherbrooke, C.C., 1986. Vari-metric: Improved approximations for multi-indenture, multi-echelon availability models. Operat. Res., 34: 311-319.
    Direct Link

  9. Sherbrooke, C.C., 2004. Optimal Inventory Modeling of Systems: Multi-Echelon Technique. 2nd Edn., Kluwer Academic Publishers, USA., ISBN-13: 978-1402078491, Pages: 368.

  10. Axsater, S., 1990. Modelling emergency lateral transshipments in inventory systems. Manage. Sci., 6: 1329-1338.
    CrossRefDirect Link

  11. Zamperini, M.B. and M. Freimer, 2005. A simulation analysis of the Vari-Metric repairable inventory optimization procedure for the US Coast Guard. Proceedings of the Winter Simulation Conference, December 4, 2005, Orlando, FL, USA., pp: 1692-1698.
    CrossRef

  12. Wong, H., G.J. van Houtum, D. Cattrysse and D. van Oudheusden, 2006. Multi-item spare parts systems with lateral transshipments and waiting time constraints. Eur. J. Operational Res., 171: 1071-1093.
    CrossRef

  13. Sun, J.S., S.J. Li, Y.M. Lv and F.G. Zhao, 2008. Simulation and research on the three-echelon storage model of the valuable spare parts in weapon equipment. Acta Armamentari, 29: 854-858 (In Chinese).
    Direct Link

  14. Wang, N. and R. Kang, 2009. Optimization of multi-echelon repairable item inventory systems with fill rate as objective. Acta Aeronau. et Astronaut. Sinica, 30: 1043-1047, (In Chinese).
    Direct Link

  15. Rui, L., K. Rui, Z.Z. Ying and H.Z. Dong, 2011. Prediction model for allocation efficiency of support equipment in products developing phase. Syst. Eng. Electronics, 33: 1040-1044.

Leave a Comment


Your email address will not be published. Required fields are marked *

Useful Links

  • Journals
  • For Authors
  • For Referees
  • For Librarian
  • For Socities

Contact Us

Office Number 1128,
Tamani Arts Building,
Business Bay,
Deira, Dubai, UAE

Phone: +971 507 888 742
Email: [email protected]

About Science Alert

Science Alert is a technology platform and service provider for scholarly publishers, helping them to publish and distribute their content online. We provide a range of services, including hosting, design, and digital marketing, as well as analytics and other tools to help publishers understand their audience and optimize their content. Science Alert works with a wide variety of publishers, including academic societies, universities, and commercial publishers.

Follow Us
© Copyright Science Alert. All Rights Reserved