American Institute of Mathematical Sciences

• Previous Article
Performance analysis of a cooperative flow game algorithm in ad hoc networks and a comparison to Dijkstra's algorithm
• JIMO Home
• This Issue
• Next Article
Global and local advertising strategies: A dynamic multi-market optimal control model
July  2019, 15(3): 1049-1083. doi: 10.3934/jimo.2018085

Optimization of a condition-based duration-varying preventive maintenance policy for the stockless production system based on queueing model

 College of information engineering, Shenzhen University, Shenzhen 518060, China

* Corresponding author: Jianyu Cao

Received  April 2016 Revised  March 2018 Published  July 2018

Fund Project: This work is supported by the National Natural Science Foundations of China (Grant Nos. 61401286, 61702341 and 61771319) and the Research Project of Shenzhen Technology University (Grant No. 201727)

A stockless production system is considered, in which the products are not produced until the orders are accepted. Due to this character, the duration of the preventive maintenance has an influence on the lead time. In addition, in this stockless production system, the cost of the preventive maintenance depends on its duration; if the lead time exceeds to the quoted lead time, some penalty cost should be considered; and the non-conforming products can still be sold by a discount. A condition-based duration-varying preventive maintenance policy is designed for the stockless production system, by making a tradeoff among the duration of the preventive maintenance, the time for the machine continuously producing, and the lead time of the order. According to the characters of the stockless production system with the designed preventive maintenance policy, it can be modeled by a BMAP/G/1 infinite-buffer queueing model with gated service and queue-length dependent vacation. Based on this queueing model, the stationary probability distributions of four performance measures for the stockless production system are analyzed, including, the number of the products produced in a production cycle, the number of the unfulfilled orders at arbitrary time, the time required to fulfill the tasks present at arbitrary time, and the lead time of the order accepted at arbitrary time. Moreover, based on some information of these performance measures, a profit function, which represents the average profit of the manufacture in a production cycle, is constructed to optimize the designed preventive maintenance policy according to specific conditions. Finally, given an example with the purchasers having different sensitivities to the lead time, some numerical experiments are carried out; and from the numerical experiments, some general results can be inferred for the stockless production system with the designed preventive maintenance policy.

Citation: Jianyu Cao, Weixin Xie. Optimization of a condition-based duration-varying preventive maintenance policy for the stockless production system based on queueing model. Journal of Industrial & Management Optimization, 2019, 15 (3) : 1049-1083. doi: 10.3934/jimo.2018085
References:

show all references

References:
${{K''}_t}\left\{ {\left. {\left( {i,v} \right) \times [0,x]} \right|\left( {{i_0},{v_0}} \right)} \right\}$ for ${{i}_{0}} = 0$, provided that at time $t$ $\left( t<{{T}_{1}} \right)$, a virtual customer arrives at the queue, while the server does not attend to the queue.
${{K''}_t}\left\{ {\left. {\left( {i,v} \right) \times [0,x]} \right|\left( {{i_0},{v_0}} \right)} \right\}$ for ${{i}_{0}}\ge 1$, provided that at time $t$ $\left( t<{{T}_{1}} \right)$, a virtual customer arrives at the queue, while the server does not attend to the queue.
${K''_t}\left\{ {\left. {\left( {i,v} \right) \times [0,x]} \right|\left( {{i_0},{v_0}} \right)} \right\}$ for $1\le {{i}_{0}}\le i$, provided that at time $t$ $\left( t<{{T}_{1}} \right)$, a virtual customer arrives at the queue, while the server is attending to the queue.
${K''_t}\left\{ {\left. {\left( {i,v} \right) \times [0,x]} \right|\left( {{i_0},{v_0}} \right)} \right\}$ for ${{i}_{0}}\ge 2$ and ${{i}_{0}}>i>0$, provided that at time $t$ $\left( t<{{T}_{1}} \right)$, a virtual customer arrives at the queue, while the server is attending to the queue.
Schematic diagrams of ${\varphi }(n)$, ${M_{c}}(x)$ and ${W_{c}}(x)$.
$\eta(q)$ vs. $q$ for the purchasers with different sensitivities to the lead time.
${{\sigma }_{m}}$ vs. $q$.
The correspondence between performance measures of the stockless production system and the queueing model
 The stockless production system The queueing model The number of the products produced in a production cycle The queue length just after the server travels to the queue The number of the unfulfilled orders at arbitrary time The queue length (including the customer in service) at arbitrary time The time required to fulfill the tasks 1 present at arbitrary time The virtual waiting time2 at arbitrary time The lead time of the order accepted at arbitrary time The actual waiting time3 at arbitrary time 1 The tasks contain the future or remaining machine set-up, machine close-down as well as PM (or the idle period) in the current production cycle, and the present unfulfilled orders.2 The definition of the virtual waiting time is given in Section 5.3 The definition of the actual waiting time is given in Section 6.
 The stockless production system The queueing model The number of the products produced in a production cycle The queue length just after the server travels to the queue The number of the unfulfilled orders at arbitrary time The queue length (including the customer in service) at arbitrary time The time required to fulfill the tasks 1 present at arbitrary time The virtual waiting time2 at arbitrary time The lead time of the order accepted at arbitrary time The actual waiting time3 at arbitrary time 1 The tasks contain the future or remaining machine set-up, machine close-down as well as PM (or the idle period) in the current production cycle, and the present unfulfilled orders.2 The definition of the virtual waiting time is given in Section 5.3 The definition of the actual waiting time is given in Section 6.

2018 Impact Factor: 1.025