Skip to main content
Log in

Optimal two-threshold replacement policy in a cumulative damage model

  • SI.: Reliability Management and Computing
  • Published:
Annals of Operations Research Aims and scope Submit manuscript

Abstract

In this paper, a two-unit system with failure interactions is studied. The system is subject to two types of shocks (I and II) and the probabilities of these two shock types are age-dependent. A type I shock, rectified by a minimal repair, causes a minor failure of unit A and type II shock causes a complete system failure that calls for a replacement. Each unit A minor failure also results in an amount of damage to unit B. The damages to unit B caused by type I shocks can be accumulated to trigger a preventive replacement or a corrective replacement action. Besides, unit B with cumulative damage of level z may become minor failed with probability \(\pi (z)\) at each unit A minor failure and rectified by a minimal repair. We consider a more general replacement policy. Under this policy, the system is preventively replaced at the Nth type I shock, or at the time when the total damage to unit B exceeds a pre-specified level Z (but less than a failure level K where \(K>Z\)) or is replaced correctively at first type II shock or when the total damage to unit B exceeding a failure level K, whichever occurs first. To minimize the expected cost per unit time, the optimal replacement policy \((N^{*}\), \(Z^{*})\) is derived analytically and determined numerically. We also show that several previous maintenance models in the literature are special cases of our model.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

References

  • Barlow, R. E., & Proschan, F. (1965). Mathematical theory of reliability. New York: Wiley.

    Google Scholar 

  • Barlow, R. E., & Proschan, F. (1975). Statistical theory of reliability. New York: Holt, Rinehart and Winston.

    Google Scholar 

  • Beichelt, F. (1993). A unifying treatment of replacement policies with minimal repair. Naval Research Logistics, 40, 51–67.

    Article  Google Scholar 

  • Block, H. W., Borges, W. S., & Savits, T. H. (1985). Age-dependent minimal repair. Journal of Applied Probability, 22(2), 370–385.

    Article  Google Scholar 

  • Boland, P. J., & Proschan, F. (1982). Periodic replacement with increasing minimal repair costs at failure. Operations Research, 30(6), 1183–1189.

    Article  Google Scholar 

  • Chang, C. C., Sheu, S. H., Chen, Y. L., & Zhang, Z. G. (2011). A multi-criteria optimal replacement policy for a system subject to shock. Computer and Industrial Engineering, 61(4), 1035–1043.

    Article  Google Scholar 

  • Chen, Y. L. (2012). A bivariate optimal imperfect preventive maintenance policy for a used system with two-type shocks. Computers & Industrial Engineering, 63(4), 1227–1234.

    Article  Google Scholar 

  • Chien, Y. H., Sheu, S. H., & Zhang, Z. G. (2012). Optimal maintenance policy for a system subject to damage in a discrete time process. Reliability Engineering and System Safety, 103, 1–10.

    Article  Google Scholar 

  • Cléroux, R., Dubuc, S., & Tilquin, C. (1979). The age replacement problem with minimal repair and random repair cost. Operations Research, 27(6), 1158–1167.

    Article  Google Scholar 

  • Cox, D. R. (1962). Renewal theory. London: Methuen.

    Google Scholar 

  • Drinkwater, R. W., & Hastings, N. A. J. (1967). A economic replacement model. Operational Research Quarterly, 18, 121–138.

    Article  Google Scholar 

  • Esary, J. D., Marshall, A. W., & Proschan, F. (1973). Shock models and wear processes. The Annals of Probability, 1(4), 627–649.

    Article  Google Scholar 

  • Gottlieb, G. (1980). Failure distributions of shock models. Journal of Applied Probability, 17(3), 745–752.

    Article  Google Scholar 

  • Hastings, N. A. J. (1969). The repair limit replacement method. Operational Research Quarterly, 20, 337–349.

    Article  Google Scholar 

  • Ito, K., & Nakagawa, T. (2011). Comparison of three cumulative damage models. Quality Technology and Quantitative Management, 8(1), 57–66.

    Article  Google Scholar 

  • Lai, M. T., & Chen, Y. C. (2006). Optimal periodic replacement policy for a two-unit system with failure rate. The International Journal of Advanced Manufacturing Technology, 29(3–4), 367–371.

    Article  Google Scholar 

  • Lai, M. T., & Chen, Y. C. (2008). Optimal replacement period of a two-unit system with failure rate interaction and external shocks. International Journal of System Science, 39(1), 71–79.

    Article  Google Scholar 

  • Makabe, H., & Morimura, H. (1965). Some considerations on preventive maintenance policies with numerical analysis. Journal of the Operations Research Society of Japan, 7, 154–171.

    Google Scholar 

  • Montoro-Cazorla, D., & Pérez-Ocón, R. (2011). Two shock and wear systems under repair standing a finite number of shocks. European Journal of Operational Research, 214(2), 298–307.

    Article  Google Scholar 

  • Murthy, D. N. P., & Nguyen, D. G. (1985). Study of two-component system with failure interaction. Naval Research Logistics, 10, 239–247.

    Article  Google Scholar 

  • Nakagawa, T. (1981). Generalized models for determining optimal number of minimal repairs before replacement. Journal of the Operations Research Society of Japan, 24, 325–337.

    Google Scholar 

  • Nakagawa, T. (2011). Stochastic processes with applications to reliability theory. Springer Series in Reliability Engineering.

  • Nakagawa, T., & Kijima, M. (1989). Replacement policies for a cumulative damage model with minimal repair at failure. IEEE Transactions on Reliability, 38(5), 581–584.

    Article  Google Scholar 

  • Nakagawa, T., & Kowada, M. (1983). Analysis of a system with minimal repair and its application to replacement policy. European Journal of Operational Research, 12, 176–182.

    Article  Google Scholar 

  • Nakagawa, T., & Murthy, D. N. P. (1993). Optimal replacement policies for a two-unit system with failure interactions. RAIRO Operations Research, 27(4), 427–438.

    Google Scholar 

  • Park, K. S. (1987). Optimal number of minor failure failures before replacement. International Journal of Systems Science, 18(2), 333–337.

    Article  Google Scholar 

  • Qian, C., Nakamura, S., & Nakagawa, T. (1999). Cumulative damage model with two kinds of shocks and its application to the backup policy. Journal of the Operations Research Society of Japan, 42(4), 501–511.

    Article  Google Scholar 

  • Qian, C., Nakamura, S., & Nakagawa, T. (2003). Replacement and minimal repair policies for a cumulative damage model with maintenance. Computers and Mathematics with Applications, 46(7), 1111–1118.

    Article  Google Scholar 

  • Qian, C., Ito, K., & Nakagawa, T. (2005). Optimal preventive maintenance policies for a shock model with given damage level. Journal of Quality in Maintenance Engineering, 11(3), 216–227.

    Article  Google Scholar 

  • Satow, T., & Osaki, S. (2003). Optimal replacement policies for a two-unit system with shock damage interaction. Computers and Mathematics with Applications, 46(7), 1129–1138.

    Article  Google Scholar 

  • Savits, T. H. (1988). Some multivariate distributions derived from a non-fatal shock model. Journal of Applied Probability, 25, 383–390.

    Article  Google Scholar 

  • Sheu, S. H., & Griffith, W. S. (1991). Multivariate age-dependent imperfect repair. Naval Research Logistics, 38, 839–850.

    Article  Google Scholar 

  • Sheu, S. H., & Griffith, W. S. (1992). Multivariate imperfect repair model. Journal Applied Probability, 29, 947–956.

    Article  Google Scholar 

  • Sheu, S. H. (1993). A generalized model for determining optimal number of minimal repairs before replacement. European Journal of Operational Research, 69(38–49), 1993.

    Google Scholar 

  • Sheu, S. H., Griffith, W. S., & Nakagawa, T. (1995). Extended optimal replacement model with random minimal repair costs. European Journal of Operational Research, 85, 636–649.

    Article  Google Scholar 

  • Sheu, S. H., Li, S. H., & Chang, C. C. (2012a). A generalized maintenance policy with age-dependent minimal repair cost for a system subject to shocks under periodic overhaul. International Journal of Systems Science, 43, 1007–1013.

    Article  Google Scholar 

  • Sheu, S. H., Chang, C. C., Zhang, Z. G., & Chien, Y. H. (2012b). A note on replacement policy for a system subject to non-homogeneous pure birth shocks. European Journal of Operational Research, 2016(2), 503–508.

    Article  Google Scholar 

  • Sheu, S. H., Chen, Y. L., Chang, C. C., & Zhang, Z. G. (2013a). Optimal number of repairs before replacement for a system subject to shocks of a non-homogeneous pure birth process. IEEE Transaction on Reliability, 62(1), 73–81.

    Article  Google Scholar 

  • Sheu, S. H., & Zhang, Z. G. (2013b). An optimal age replacement policy for multi-state systems. IEEE Transaction on Reliability, 62(3), 722–735.

    Article  Google Scholar 

  • Sheu, S. H., Sung, C. K., Hsu, T. S., & Chen, Y. C. (2013c). Age replacement policy for a two-unit system subject to non-homogeneous pure birth shocks. Applied Mathematical Modelling, 37(10–11), 7027–7036.

    Article  Google Scholar 

  • Sung, C. K., Sheu, S. H., Hsu, T. S., & Chen, Y. C. (2013). Extended optimal replacement policy for a two-unit system with failure rate interaction and external shocks. International Journal of Systems Science, 44(5), 877–888.

    Article  Google Scholar 

  • Zhao, X., & Nakagawa, T. (2012). Optimization problems of replacement first or last in reliability theory. European Journal of Operational Research, 223(1), 141–149.

    Article  Google Scholar 

  • Zhao, X., Qian, C., & Nakagawa, T. (2013). Optimal policies for cumulative damage models with maintenance last and first. Reliability Engineering and System Safety, 110, 50–59.

    Article  Google Scholar 

  • Zhao, X., Nakagawa, T., & Qian, C. (2012). Optimal imperfect preventive maintenance policies for a used system. International Journal of Systems Science, 43(9), 1632–1641.

    Article  Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Corresponding authors

Correspondence to Shey-Huei Sheu or Zhe-George Zhang.

Appendix

Appendix

1.1 Proof of (6) + (7) + (8) + (9) = 1

$$\begin{aligned}&P_N +P_Z +P_K +P_{II}\\&\quad = G^{\left( N \right) }\left( Z \right) \int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_{N-1} (t)dt} +\sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {P_j (t)dF_P (t)} }\\&\qquad + \sum _{j=0}^{N-1} {\int _0^Z {\left[ {G_{j+1} \left( {K-y} \right) -G_{j+1} \left( {Z-y} \right) } \right] dG^{\left( j \right) }\left( y \right) } \times \int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_j (t)dt} }\\&\qquad +\sum _{j=0}^{N-1} {\int _0^Z {\bar{{G}}_{j+1} \left( {K-y} \right) dG^{(j)}\left( y \right) } \times \int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_j (t)dt}}\\&\quad = G^{\left( N \right) }\left( Z \right) \int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_{N-1} (t)dt} +\sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {P_j (t)dF_P (t)} }\\&\qquad +\sum _{j=0}^{N-1} {\left[ {G^{(j)}\left( Z \right) -G^{(j+1)}\left( Z \right) } \right] \int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_j (t)dt} }\\&\quad = G^{\left( N \right) }\left( Z \right) \int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_{N-1} (t)dt} +\sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_j (t)dt}}\\&\qquad -\sum _{j=1}^N {G^{(j)}\left( Z \right) \int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_{j-1} (t)dt} } +\sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {P_j (t)dF_P (t)}}\\ \end{aligned}$$
$$\begin{aligned}&\quad \!= \!\sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_j (t)dt} } \!-\!\sum _{j=1}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_{j-1} (t)dt}}\\&\qquad + \sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {P_j (t)dF_P (t)}}\\&\quad = \sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {P_j (t)dF_P (t)} } -\sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {{\bar{F}}_P (t)dP_j (t)}}\\&\quad = \sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {P_j (t)dF_P (t)} } -\sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \left[ {\int _0^\infty {P_j (t)dF_P (t)} -{\bar{F}}_P \left( 0 \right) P_j \left( 0 \right) } \right] }\\&\quad = 1 \end{aligned}$$

1.2 Derivation of (12)

$$\begin{aligned} E\left( {U_1 } \right)= & {} G^{(N)}\left( Z \right) \int _0^\infty {t{\bar{F}}_P (t)q(t)r(t)P_{N-1} (t)dt} +\sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {tP_j (t)dF_P (t)} }\\&+ \sum _{j=0}^{N-1} \int _0^Z {\left[ {G_{j+1} \left( {K-y} \right) -G_{j+1} \left( {Z-y} \right) } \right] dG^{\left( j \right) }\left( y \right) }\\&\times \int _0^\infty {t{\bar{F}}_P (t)q(t)r(t)P_j (t)dt}\\&+ \sum _{j=0}^{N-1} {\int _0^Z {\bar{{G}}_{j+1} \left( {K-y} \right) dG^{(j)}\left( y \right) } \times \int _0^\infty {t{\bar{F}}_P (t)q(t)r(t)P_j (t)dt} }\\ \!= & {} \!G^{(N)}\left( Z \right) \int _0^\infty {t{\bar{F}}_P (t)q(t)r(t)P_{N-1} (t)dt} +\sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {tP_j (t)dF_P (t)} }\\&+ \sum _{j=0}^{N-1} {\left[ {G^{(j)}\left( Z \right) -G^{(j+1)}\left( Z \right) } \right] \int _0^\infty {t{\bar{F}}_P (t)q(t)r(t)P_j (t)dt} }\\= & {} G^{(N)}\left( Z \right) \int _0^\infty {t{\bar{F}}_P (t)q(t)r(t)P_{N-1} (t)dt}\\&+ \sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {t{\bar{F}}_P (t)q(t)r(t)P_j (t)dt}}\\&-\sum _{j=1}^N {G^{(j)}\left( Z \right) \int _0^\infty {t{\bar{F}}_P (t)q(t)r(t)P_{j-1} (t)dt} } \!+\!\sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {tP_j (t)dF_P (t)}}\\ \end{aligned}$$
$$\begin{aligned}= & {} \sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {t{\bar{F}}_P (t)q(t)r(t)P_j (t)dt}}\\&-\sum _{j=1}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {t{\bar{F}}_P (t)q(t)r(t)P_{j-1} (t)dt} } +\sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {tP_j (t)dF_P (t)}}\\= & {} -\sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {t{\bar{F}}_P (t)dP_j (t)} } +\sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {tP_j (t)dF_P (t)}}\\= & {} \sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {{\bar{F}}_P (t)P_j (t)dt} } -\sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {tP_j (t)dF_P (t)}}\\&+\sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {tP_j (t)dF_P (t)} }\\= & {} \sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {{\bar{F}}_P (t)P_j (t)dt} } \end{aligned}$$

1.3 Derivation of (14)

$$\begin{aligned} E\left( {R_1 } \right)= & {} c_1 P_N +c_2 P_Z +c_3 P_{II} +c_4 P_K +\sum _{j=1}^{N-1} {G^{\left( j \right) }\left( Z \right) \int _0^\infty {\alpha _j (t){\bar{F}}_P (t)q(t)r(t)P_{j-1} (t)dt}}\\&+\sum _{j=1}^{N-1} {\int _0^Z {\pi \left( y \right) \eta \left( y \right) dG^{\left( j \right) }\left( y \right) } \times \int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_{j-1} (t)dt}}\\= & {} c_1 G^{\left( N \right) }\left( Z \right) \int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_{N-1} (t)dt} +c_2 \left( {1-P_N -P_{II} -P_K } \right) \\&+\, c_3 \sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {P_j (t)dF_P (t)}}\\&+\, c_4 \sum _{j=0}^{N-1} {\int _0^Z {\bar{{G}}_{j+1} \left( {K-y} \right) dG^{(j)}\left( y \right) } \times \int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_j (t)dt}}\\&+ \sum _{j=1}^{N-1} {G^{\left( j \right) }\left( Z \right) \int _0^\infty {\alpha _j (t){\bar{F}}_P (t)q(t)r(t)P_{j-1} (t)dt}}\\&+ \sum _{j=1}^{N-1} {\int _0^Z {\pi \left( y \right) \eta \left( y \right) dG^{\left( j \right) }\left( y \right) } \times \int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_{j-1} (t)dt}}\\ \end{aligned}$$
$$\begin{aligned}= & {} c_2 +\left( {c_1 -c_2 } \right) G^{\left( N \right) }\left( Z \right) \int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_{N-1} (t)dt}\\&+ \left( {c_3 -c_2 } \right) \sum _{j=0}^{N-1} {G^{(j)}\left( Z \right) \int _0^\infty {P_j (t)dF_P (t)}}\\&+ \left( {c_4 -c_2 } \right) \sum _{j=0}^{N-1} {\int _0^Z {\bar{{G}}_{j+1} \left( {K-y} \right) dG^{(j)}\left( y \right) } \times \int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_j (t)dt}}\\&+ \sum _{j=1}^{N-1} {G^{\left( j \right) }\left( Z \right) \int _0^\infty {\alpha _j (t){\bar{F}}_P (t)q(t)r(t)P_{j-1} (t)dt}}\\&+ \sum _{j=1}^{N-1} {\int _0^Z {\pi \left( y \right) \eta \left( y \right) dG^{\left( j \right) }\left( y \right) } \times \int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_{j-1} (t)dt} } \end{aligned}$$

1.4 Proof of Lemma 1-(i)

First, we rewrite \(A_N \) as follows.

$$\begin{aligned} A_N= & {} \frac{G^{\left( N \right) }\left( Z \right) \int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_{N-1} (t)dt} -G^{\left( {N+1} \right) }\left( Z \right) \int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_N (t)dt} }{G^{\left( N \right) }\left( Z \right) \int _0^\infty {{\bar{F}}_P (t)P_N (t)dt}}\\= & {} \left[ {1-\frac{G^{\left( {N+1} \right) }\left( Z \right) }{G^{\left( N \right) }\left( Z \right) }+\frac{G^{\left( {N+1} \right) }\left( Z \right) }{G^{\left( N \right) }\left( Z \right) }} \right] \frac{\int _0^\infty {{\bar{F}}_P (t)r(t)P_N (t)dt} }{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt}}\\&-\frac{G^{\left( {N+1} \right) }\left( Z \right) \int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_N (t)dt} }{G^{\left( N \right) }\left( Z \right) \int _0^\infty {{\bar{F}}_P (t)P_N (t)dt}}\\= & {} \left[ {1-\frac{G^{\left( {N+1} \right) }\left( Z \right) }{G^{\left( N \right) }\left( Z \right) }} \right] \left[ {\frac{\int _0^\infty {{\bar{F}}_P (t)r(t)P_N (t)dt} }{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt} }} \right] \\&-\left\{ {\frac{G^{\left( {N+1} \right) }\left( Z \right) }{G^{\left( N \right) }\left( Z \right) }\left[ {-\frac{\int _0^\infty {{\bar{F}}_P (t)p(t)r(t)P_N (t)dt} }{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt} }} \right] } \right\} \end{aligned}$$

We first proof \({\int _0^\infty {{\bar{F}}_P (t)r(t)P_N (t)dt} }/{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt} }\) is increasing in N.

Let

$$\begin{aligned} J_1 (T)= & {} \int _0^T {{\bar{F}}_P (t)r(t)P_{N+1} (t)dt} \int _0^T {{\bar{F}}_P (t)P_N (t)dt}\\&-\int _0^T {{\bar{F}}_P (t)r(t)P_N (t)dt} \int _0^T {{\bar{F}}_P (t)P_{N+1} (t)dt} \end{aligned}$$

Differentiating \(J_1 (T)\) with respect to T, then we obtain

$$\begin{aligned} J_1^{\prime } (T)= & {} \int _0^T {{\bar{F}}_P (T)r(T)P_{N+1} (T){\bar{F}}_P (t)P_N (t)dt} +\int _0^T {{\bar{F}}_P (t)r(t)P_{N+1} (t){\bar{F}}_P (T)P_N (T)dt}\\&-\int _0^T {{\bar{F}}_P (T)r(T)P_N (T){\bar{F}}_P (t)P_{N+1} (t)dt} -\int _0^T {{\bar{F}}_P (t)r(t)P_N (t){\bar{F}}_P (T)P_{N+1} (T)dt}\\= & {} {\bar{F}}_P (T)\int _0^T {{\bar{F}}_P (t)\left[ {r(T)-r(t)} \right] \left[ {P_{N+1} (T)P_N (t)-P_{N+1} (t)P_N (T)} \right] dt} >0 \end{aligned}$$

since r(t) is increasing with t, we obtain \(J_1 (0)=0\) and \(J_1^{\prime } (T)>0\). Hence, \(J_1 (T)>0\) for all \(T>0\) and \({\int _0^\infty {{\bar{F}}_P (t)r(t)P_N (t)dt} }/{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt}}\) is increasing in N.

Evidently, for and N,

$$\begin{aligned} \frac{\int _0^\infty {{\bar{F}}_P (t)r(t)P_N (t)dt} }{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt} }\le r\left( \infty \right) \end{aligned}$$

For any \(T\in \left( {0,\infty } \right) \), we have

$$\begin{aligned}&\frac{\int _0^\infty {{\bar{F}}_P (t)r(t)P_N (t)dt} }{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt} }\ge \frac{\int _T^\infty {{\bar{F}}_P (t)r(t)P_N (t)dt} }{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt} }\\&\quad \ge \frac{r(T)\int _T^\infty {{\bar{F}}_P (t)P_N (t)dt} }{\int _0^T {{\bar{F}}_P (t)P_N (t)dt} +\int _T^\infty {{\bar{F}}_P (t)P_N (t)dt} }\ge \frac{r(T)}{1+\left\{ {\frac{\int _0^T {{\bar{F}}_P (t)P_N (t)dt} }{\int _T^\infty {{\bar{F}}_P (t)P_N (t)dt} }} \right\} } \end{aligned}$$

Further, the bracket of the denominator is, for \(T_1 \in \left( {T,\infty } \right) \),

$$\begin{aligned} \frac{\int _0^T {{\bar{F}}_P (t)P_N (t)dt} }{\int _T^\infty {{\bar{F}}_P (t)P_N (t)dt} }\le & {} \frac{\int _0^T {\exp \left\{ {-\int _0^t {r(x)dx} } \right\} \left( {\int _0^t {q(x)r(x)dx} } \right) ^{N}dt} }{\int _{T_1 }^\infty {\exp \left\{ {-\int _0^t {r(x)dx} } \right\} \left( {\int _0^t {q(x)r(x)dx} } \right) ^{N}dt} }\\\le & {} \frac{\int _0^T {\exp \left\{ {-\int _0^t {r(x)dx} } \right\} dt} }{\left( {\frac{\int _0^{T_1 } {q(x)r(x)dx} }{\int _0^T {q(x)r(x)dx} }} \right) ^{N}\int _{T_1 }^\infty {\exp \left\{ {-\int _0^t {r(x)dx} } \right\} dt} }\rightarrow 0\hbox { as }N\rightarrow \infty \end{aligned}$$

Thus,

$$\begin{aligned} r\left( \infty \right) =\mathop {\lim }\limits _{N\rightarrow \infty } r\left( \infty \right) \ge \mathop {\lim }\limits _{N\rightarrow \infty } \frac{\int _0^\infty {{\bar{F}}_P (t)r(t)P_N (t)dt} }{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt} }\ge r(T) \end{aligned}$$

Since T is arbitrary in \(\left( {0,\infty } \right) \), we have

$$\begin{aligned} \mathop {\lim }\limits _{N\rightarrow \infty } \frac{\int _0^\infty {{\bar{F}}_P (t)r(t)P_N (t)dt} }{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt} }=r\left( \infty \right) \end{aligned}$$

From the above result and Lemma 1-(ii), we can show that \(A_N\) is increasing in N, and diverges to \(\infty \) as \(N \rightarrow \infty \) if \(G^{(N)}(\cdot )\) is non-increasing in N and \({G^{({N+1})}(\cdot )}/{G^{(N)}(\cdot )}\) is non-increasing in N.

1.5 Proof of Lemma 1-(ii)

Let

$$\begin{aligned} J_2 (T)= & {} \int _0^T {{\bar{F}}_P (t)p(t)r(t)P_{N+1} (t)dt} \int _0^T {{\bar{F}}_P (t)P_N (t)dt}\\&-\int _0^T {{\bar{F}}_P (t)p(t)r(t)P_N (t)dt} \int _0^T {{\bar{F}}_P (t)P_{N+1} (t)dt} \end{aligned}$$

Differentiating \(J_2 (T)\) with respect to T, then we obtain

$$\begin{aligned} J_2^{\prime } (T)= & {} \int _0^T {{\bar{F}}_P (T)p(T)r(T)P_{N+1} (T){\bar{F}}_P (t)P_N (t)dt}\\&+\int _0^T {{\bar{F}}_P (t)p(t)r(t)P_{N+1} (t){\bar{F}}_P (T)P_N (T)dt}\\&-\int _0^T {{\bar{F}}_P (T)p(T)r(T)P_N (T){\bar{F}}_P (t)P_{N+1} (t)dt}\\&-\int _0^T {{\bar{F}}_P (t)p(t)r(t)P_N (t){\bar{F}}_P (T)P_{N+1} (T)dt}\\= & {} {\bar{F}}_P (T)\int _0^T {\bar{F}}_P (t)\left[ {p(T)r(T)-p(t)r(t)} \right] \left[ P_{N+1} (T)P_N (t)\right. \\&\left. -P_{N+1} (t)P_N (T) \right] dt >0 \end{aligned}$$

since r(t) is increasing with t and p(t) is non-decreasing with t, we obtain \(J_2 \left( 0 \right) =0\) and \(J_2^{\prime } (T)>0\). Hence, \(J_2 (T)>0\) for all \(T>0\) and \(B_N \) is increasing in N.

Evidently, for and N,

$$\begin{aligned} \frac{\int _0^\infty {{\bar{F}}_P (t)p(t)r(t)P_N (t)dt} }{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt} }\le p\left( \infty \right) r\left( \infty \right) \end{aligned}$$

For any \(T\in \left( {0,\infty } \right) \), we have

$$\begin{aligned}&\frac{\int _0^\infty {{\bar{F}}_P (t)p(t)r(t)P_N (t)dt} }{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt} }\ge \frac{\int _T^\infty {{\bar{F}}_P (t)p(t)r(t)P_N (t)dt} }{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt} }\\&\ge \frac{p(T)r(T)\int _T^\infty {{\bar{F}}_P (t)P_N (t)dt} }{\int _0^T {{\bar{F}}_P (t)P_N (t)dt} +\int _T^\infty {{\bar{F}}_P (t)P_N (t)dt} }\ge \frac{p(T)r(T)}{1+\left\{ {\frac{\int _0^T {{\bar{F}}_P (t)P_N (t)dt} }{\int _T^\infty {{\bar{F}}_P (t)P_N (t)dt} }} \right\} } \end{aligned}$$

where \({\int _0^T {{\bar{F}}_P (t)P_N (t)dt} }/{\int _T^\infty {{\bar{F}}_P (t)P_N (t)dt} }\) is as given in the above and has the same property. Thus,

$$\begin{aligned} p\left( \infty \right) r\left( \infty \right) =\mathop {\lim }\limits _{N\rightarrow \infty } p\left( \infty \right) r\left( \infty \right) \ge \mathop {\lim }\limits _{N\rightarrow \infty } \frac{\int _0^\infty {{\bar{F}}_P (t)p(t)r(t)P_N (t)dt} }{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt} }\ge p(T)r(T) \end{aligned}$$

Because T is arbitrary in \(\left( {0,\infty } \right) \), we have \(\lim \limits _{N\rightarrow \infty } B_N =p\left( \infty \right) r\left( \infty \right) \).

1.6 Proof of Lemma 1-(iii)

Using the skills of the integration by parts, we have

$$\begin{aligned} C_N= & {} \frac{\int _0^Z {\bar{{G}}_{N+1} \left( {K-y} \right) dG^{\left( N \right) }\left( y \right) } }{G^{\left( N \right) }\left( Z \right) }\frac{\int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_N (t)dt} }{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt}}\\= & {} \frac{\int _0^Z {\bar{{G}}_{N+1} \left( {K-y} \right) dG^{\left( N \right) }\left( y \right) } }{G^{\left( N \right) }\left( Z \right) }\frac{\int _0^\infty {{\bar{F}}_P (t)r(t)P_{N+1} (t)dt} }{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt}} \end{aligned}$$

Let

$$\begin{aligned} J_3 (T)= & {} \int _0^T {{\bar{F}}_P (t)r(t)P_{N+2} (t)dt} \hbox { }\times \int _0^T {{\bar{F}}_P (t)P_N (t)dt}\\&-\int _0^T {{\bar{F}}_P (t)r(t)P_{N+1} (t)dt} \quad \times \int _0^T {{\bar{F}}_P (t)P_{N+1} (t)dt} \end{aligned}$$

Differentiating \(J_3 (T)\) with respect to T, then we obtain

$$\begin{aligned} J_3 (T)= & {} \int _0^T {{\bar{F}}_P (T)r(T)P_{N+2} (T){\bar{F}}_P (t)P_N (t)dt} +\int _0^T {{\bar{F}}_P (t)r(t)P_{N+2} (t){\bar{F}}_P (T)P_N (T)dt}\\&-\int _0^T {{\bar{F}}_P (T)r(T)P_{N+1} (T){\bar{F}}_P (t)P_{N+1} (t)dt}\\&-\int _0^T {{\bar{F}}_P (t)r(t)P_{N+1} (t){\bar{F}}_P (T)P_{N+1} (T)dt}\\= & {} \int _0^T {{\bar{F}}_P (T)r(T){\bar{F}}_P (t)\left[ {P_{N+2} (T)P_N (t)-P_{N+1} (T)P_{N+1} (t)} \right] dt}\\&+\int _0^T {{\bar{F}}_P (T)r(t){\bar{F}}_P (t)\left[ {P_{N+2} (t)P_N (T)-P_{N+1} (T)P_{N+1} (t)} \right] dt}\\= & {} \int _0^T {{\bar{F}}_P (T){\bar{F}}_P (t)\left[ {r(T)P_{N+1} (T)P_N (t)-r(t)P_{N+1} (t)P_{N+1} (T)} \right] }\\&\times \left[ {\frac{P_{N+2} (T)}{P_{N+1} (T)}-\frac{P_{N+2} (t)}{P_{N+1} (t)}} \right] dt\\&+\int _0^T {{\bar{F}}_P (T){\bar{F}}_P (t)r(T)P_{N+1} (T)P_N (T)\left[ {\frac{P_{N+2} (t)}{P_{N+1} (t)}-\frac{P_{N+1} (t)}{P_N (t)}} \right] dt}\\&+\int _0^T {{\bar{F}}_P (T){\bar{F}}_P (t)r(T)P_{N+1} (t)P_N (T)\left[ {\frac{P_{N+2} (T)}{P_{N+1} (T)}-\frac{P_{N+1} (T)}{P_N (T)}} \right] dt} \end{aligned}$$

since r(t) is increasing with t and p(t) is non-decreasing with t, we obtain \(J_3 (0)=0\) and \(J_3^{\prime } (T)>0\). Hence, \(J_3 (T)>0\) for all \(T>0\) and \({\int _0^\infty {{\bar{F}}_P (t)r(t)P_{N+1} (t)dt} }/{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt} }\) is increasing in N. Hence, \(C_N \) is increasing in N if \({\int _0^Z {\bar{{G}}_{N+1} ({K-y})dG^{(N)}(y)} }/{G^{(N)}(Z)}\) is non-decreasing in N.

From the above result and Lemma 1-(ii), we can show that \(\lim \limits _{N \rightarrow \infty }\) \(C_N =\lim \limits _{N \rightarrow \infty }\) \(\frac{\int _0^Z {\bar{{G}}_{N+1} ({K-y})dG^{(N)}(y)} }{G^{(N)}(Z)}\) \(\frac{\int _0^\infty {{\bar{F}}_P (t)q(t)r(t)P_N (t)dt} }{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt} }\)

$$\begin{aligned}= & {} \mathop {\lim }\limits _{N\rightarrow \infty } \frac{\int _0^Z {\bar{{G}}_{N+1} \left( {K-y} \right) dG^{\left( N \right) }\left( y \right) } }{G^{\left( N \right) }\left( Z \right) }\\&\times \mathop {\lim }\limits _{N\rightarrow \infty } \frac{\int _0^\infty {{\bar{F}}_P (t)r(t)P_N (t)dt} -\int _0^\infty {{\bar{F}}_P (t)p(t)r(t)P_N (t)dt} }{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt}}\\= & {} \mathop {\lim }\limits _{N\rightarrow \infty } \frac{\int _0^Z {\bar{{G}}_{N+1} \left( {K-y} \right) dG^{\left( N \right) }\left( y \right) } }{G^{\left( N \right) }\left( Z \right) }\\&\times \left[ {\mathop {\lim }\limits _{N\rightarrow \infty } \frac{\int _0^\infty {{\bar{F}}_P (t)r(t)P_N (t)dt} }{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt} }-\mathop {\lim }\limits _{N\rightarrow \infty } \frac{\int _0^\infty {{\bar{F}}_P (t)p(t)r(t)P_N (t)dt} }{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt} }} \right] \\= & {} \mathop {\lim }\limits _{N\rightarrow \infty } \frac{\int _0^Z {\bar{{G}}_{N+1} \left( {K-y} \right) dG^{\left( N \right) }\left( y \right) } }{G^{\left( N \right) }\left( Z \right) } \times \left[ {r\left( \infty \right) -p\left( \infty \right) r\left( \infty \right) } \right] \\= & {} \mathop {\lim }\limits _{N\rightarrow \infty } \frac{\int _0^Z {\bar{{G}}_{N+1} \left( {K-y} \right) dG^{\left( N \right) }\left( y \right) } }{G^{\left( N \right) }\left( Z \right) } \times q\left( \infty \right) r\left( \infty \right) \end{aligned}$$

1.7 Proof of Lemma 1-(iv)

Let

$$\begin{aligned} J_4 (T)= & {} \int _0^T {\left[ {\alpha _{N+1} (t)r(t)-\alpha _{N+1}^{\prime } (t)+\phi _{N+1} \left( Z \right) r(t)} \right] {\bar{F}}_P (t)P_{N+1} (t)dt}\\&\times \int _0^T {{\bar{F}}_P (t)P_N (t)dt}\\&-\int _0^T {\left[ {\alpha _N (t)r(t)-\alpha _N^{\prime } (t)+\phi _N \left( Z \right) r(t)} \right] {\bar{F}}_P (t)P_N (t)dt}\\&\times \int _0^T {{\bar{F}}_P (t)P_{N+1} (t)dt} \end{aligned}$$

Differentiating \(J_4 (T)\) with respect to T, then we obtain

$$\begin{aligned} J_4^{\prime } (T)= & {} \int _0^T {\left[ {\alpha _{N+1} \left( \hbox {T} \right) r(T)-\alpha _{N+1}^{\prime } (T)+\phi _{N+1} \left( Z \right) r(T)} \right] {\bar{F}}_P (T)P_{N+1} (T){\bar{F}}_P (t)P_N (t)dt}\\&+\int _0^T {\left[ {\alpha _{N+1} (t)r(t)-\alpha _{N+1}^{\prime } (t)+\phi _{N+1} \left( Z \right) r(t)} \right] {\bar{F}}_P (t)P_{N+1} (t){\bar{F}}_P (T)P_N (T)dt}\\&-\int _0^T {\left[ {\alpha _N (T)r(T)-\alpha _N^{\prime } (T)+\phi _N \left( Z \right) r(T)} \right] {\bar{F}}_P (T)P_N (T){\bar{F}}_P (t)P_{N+1} (t)dt}\\&-\int _0^T {\left[ {\alpha _N (t)r(t)-\alpha _N^{\prime } (t)+\phi _N \left( Z \right) r(t)} \right] {\bar{F}}_P (t)P_N (t){\bar{F}}_P (T)P_{N+1} (T)dt}\\= & {} {\bar{F}}_P (T)\int _0^T {{\bar{F}}_P (t)\left\{ {\left[ {\alpha _{N+1} \left( \hbox {T} \right) r(T)-\alpha _{N+1}^{\prime } (T)+\phi _{N+1} \left( Z \right) r(T)} \right] } \right. }\\&\left. {-\left[ {\alpha _N (t)r(t)-\alpha _N^{\prime } (t)+\phi _N \left( Z \right) r(t)} \right] } \right\} P_{N+1} (T)P_N (t)dt\\&-\left\{ {{\bar{F}}_P (T)\int _0^T {{\bar{F}}_P (t)\left\{ {\left[ {\alpha _N \left( \hbox {T} \right) r(T)-\alpha _N^{\prime } (T)+\phi _N \left( Z \right) r(T)} \right] } \right. } } \right. \\&\left. {\left. {-\left[ {\alpha _{N+1} (t)r(t)-\alpha _{N+1}^{\prime } (t)+\phi _{N+1} \left( Z \right) r(t)} \right] } \right\} P_{N+1} (t)P_N (T)dt} \right\} \\\ge & {} {\bar{F}}_P (T)\int _0^T {{\bar{F}}_P (t)\left\{ {\left[ {\alpha _{N+1} \left( \hbox {T} \right) r(T)-\alpha _{N+1}^{\prime } (T)+\phi _{N+1} \left( Z \right) r(T)} \right] } \right. }\\&\left. {-\left[ {\alpha _{N+1} (t)r(t)-\alpha _{N+1}^{\prime } (t)+\phi _{N+1} \left( Z \right) r(t)} \right] } \right\} P_{N+1} (T)P_N (t)dt\\&-\left\{ {{\bar{F}}_P (T)\int _0^T {{\bar{F}}_P (t)\left\{ {\left[ {\alpha _{N+1} \left( \hbox {T} \right) r(T)-\alpha _{N+1}^{\prime } (T)+\phi _{N+1} \left( Z \right) r(T)} \right] } \right. } } \right. \\&\left. {\left. {-\left[ {\alpha _{N+1} (t)r(t)-\alpha _{N+1}^{\prime } (t)+\phi _{N+1} \left( Z \right) r(t)} \right] } \right\} P_{N+1} (t)P_N (T)dt} \right\} \\= & {} {\bar{F}}_P (T)\int _0^T {{\bar{F}}_P (t)\left\{ {\left[ {\alpha _{N+1} \left( \hbox {T} \right) r(T)-\alpha _{N+1}^{\prime } (T)+\phi _{N+1} \left( Z \right) r(T)} \right] } \right. }\\&\left. {-\left[ {\alpha _{N+1} (t)r(t)-\alpha _{N+1}^{\prime } (t)+\phi _{N+1} \left( Z \right) r(t)} \right] } \right\} \left[ P_{N+1} (T)P_N (t)\right. \\&\left. -P_{N+1} (t)P_N (T) \right] dt\\&>0 \end{aligned}$$

since r(t) is increasing with t and \(\alpha _j (t)r(t)-\alpha _j^{\prime } (t)+\phi _j \left( Z \right) r(t)\) is non-decreasing with t, we obtain \(J_4 \left( 0 \right) =0\) and \(J_4^{\prime } (T)>0\). Hence, \(J_4 (T)>0\) for all \(T>0\) and \(D_N \) is increasing in N.

Evidently, for and N,

$$\begin{aligned} D_N= & {} \frac{\int _0^\infty {\left[ {\alpha _N (t)r(t)-\alpha _N^{\prime } (t)+\phi _N \left( Z \right) r(t)} \right] {\bar{F}}_P (t)P_N (t)dt}}{\int _0^\infty {{\bar{F}}_P (t)P_N (t)dt}}\\\le & {} \alpha _\infty \left( \infty \right) r\left( \infty \right) -\alpha _\infty ^{\prime } \left( \infty \right) +\phi _\infty \left( Z \right) r\left( \infty \right) \end{aligned}$$

For any \(T\in \left( {0,\infty } \right) \), we have

$$\begin{aligned}&\frac{\int _0^\infty {\left[ {\alpha _N (t)r(t)-\alpha _N^{\prime } (t)+\phi _N \left( Z \right) r(t)} \right] {\bar{F}}_P (t)P_N (t)dt}}{\int _0^T {{\bar{F}}_P (t)P_N (t)dt}}\\&\quad \ge \frac{\int _T^\infty {\left[ {\alpha _N (t)r(t)-\alpha _N^{\prime } (t)+\phi _N \left( Z \right) r(t)} \right] {\bar{F}}_P (t)P_N (t)dt}}{\int _0^T {{\bar{F}}_P (t)P_N (t)dt}}\\&\quad \ge \frac{\left[ {\alpha _N (T)r(T)-\alpha _N^{\prime } (T)+\phi _N \left( Z \right) r(T)} \right] \int _T^\infty {{\bar{F}}_P (t)P_N (t)dt} }{\int _0^T {{\bar{F}}_P (t)P_N (t)dt} +\int _T^\infty {{\bar{F}}_P (t)P_N (t)dt}}\\&\quad \ge \frac{\alpha _N (T)r(T)-\alpha _N^{\prime } (T)+\phi _N \left( Z \right) r(T)}{1+\left\{ {\frac{\int _0^T {{\bar{F}}_P (t)P_N (t)dt} }{\int _T^\infty {{\bar{F}}_P (t)P_N (t)dt} }} \right\} } \end{aligned}$$

where \({\int _0^T {{\bar{F}}_P (t)P_N (t)dt} }/{\int _T^\infty {{\bar{F}}_P (t)P_N (t)dt} }\) is as given in the above and has the same property. Thus,

$$\begin{aligned}&\alpha _\infty (\infty )r(\infty )-\alpha _\infty ^{\prime } (\infty )+\phi _\infty \left( Z \right) r(\infty ) \\&\quad =\mathop {\lim }\limits _{N\rightarrow \infty } [\alpha _N (\infty )r(\infty )-\alpha _N^{\prime } (\infty )+\phi _N \left( Z \right) r(\infty )]\\&\quad \ge \mathop {\lim }\limits _{N\rightarrow \infty } \frac{\int _{\hbox { }0}^{\hbox { }T} {\left[ {\alpha _N (t)r(t)-\alpha _N^{\prime } (t)+\phi _N \left( Z \right) r(t)} \right] {\bar{F}}_p (t)P_N (t)\hbox {d}t}}{\int _{\hbox { }0}^{\hbox { }T} {{\bar{F}}_p (t)P_N (t)\hbox {d}t}}\\&\quad \ge \alpha _\infty (T)r(T)-\alpha _\infty ^{\prime } (T_1)+\phi _\infty \left( Z \right) r(T). \end{aligned}$$

Because T is arbitrary in \(({0,\infty })\), we have

$$\begin{aligned} \mathop {\lim }\limits _{N\rightarrow \infty } D_N =\alpha _\infty (\infty )r(\infty )-\alpha _\infty ^{\prime } (\infty )+\phi _\infty \left( Z \right) r(\infty ). \end{aligned}$$

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Sheu, SH., Liu, TH., Zhang, ZG. et al. Optimal two-threshold replacement policy in a cumulative damage model. Ann Oper Res 244, 23–47 (2016). https://doi.org/10.1007/s10479-016-2142-3

Download citation

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10479-016-2142-3

Keywords

Navigation