Zhiyuan Rui

Near-Optimal Production and Maintenance Control for Failure Prone Production Systems with Demand Uncertainty

Under the circumstances that the demand is uncertain, furthermore, the demand sometimes cannot be satisfied by the production, the paper studies the production and maintenance control problem of failure prone production systems. For the hybrid system, a so-called double level hedging point control policy considering purchasing extra production or maintenance capacity is presented, which has strong dynamic characteristics. By the method that studies the problem from over the finite horizon to over the infinite horizon, meanwhile, utilizing a decomposition method, the corresponding hedging points are obtained respectively. The numerical comparison specifies and to some extent testifies the method.

A Scheduling Model of Unreliable Manufacturing Systems on Receding Horizon

An unreliable manufacturing system is examined. The system can produce several types of products. A setup (with setup time or cost or both) is considered when the production is switched from one type of products to another. The decision variables are a sequence of setups and the production plan. After a more straightforward necessary condition and a necessary and sufficient condition for the optimality of the just-in-time (JIT) control policy were presented respectively, based on the hedging point control policy and the idea of JIT (just-in-time), a new real-time receding horizon scheduling algorithm is presented. Numerical results are also presented.

A New Method for Computing Contact Angle of High Speed Ball Bearing

With the development of high-speed and precision of ball bearing, it was found that conventional model using the raceway control theory failed to give satisfactory model accuracy. A new calculating method of contact angle of ball bearing, that considers elastohydrodynamic lubrication of high speed ball bearing and the simultaneous spin motion between the balls and the inner and outer raceway, is proposed. Utilizing rigid ferrule theory and Hertzian contact theory, the influences of contact load on friction coefficient are analyzed. The formula of the attitude angle of rolling element based on the d’Alembert’s principle is introduced into the dynamic model. A case of contact angles is simulated with two computing methods. The comparison results show that the proposed method has much better accuracy than the conventional model and are suitable for precision bearing.

A new extended system aggregation method of hybrid production lines

Hybrid production lines with unreliable machines and buffers are examined. Different from traditional aggregation method, the unreliable buffers are originally considered and a more general aggregation method is offered, in which not only the unreliable buffers are considered, but also the probabilities of system states are obtained by a discrete model rather than the continuous flow model of unreliable manufacturing systems. The solution technique is offered to get the system sates probabilities. The method advances the traditional system aggregation techniques. Numerical results specify the extended aggregation method and also show that the unreliable limited buffers have a strong impact on the efficiency of the production lines.

Analysis of Two-machine Lines with Multiple Failure Modes and Unreliable Buffer

This paper presents an method for evaluating the performance of production lines with a finite buffer and two unreliable machines. In this system, each machine can fail in more than one way, and the buffer also can fail. We study the transitions of the states probabilities based on the discrete model, a corresponding algorithm are developed here, and an approximate technique is proposed to evaluate the performance of hybrid production systems. The experiment confirms that this method is a robust one.

The Buffer Size Control Problem for Hybrid Systems with Machines' Capacity Uncertainty

The failure prone production systems are examined. For the hybrid systems, the failure of machines is unavoidable and the production capacity of machines is uncertain. The continuous flow model of the unreliable manufacturing systems is considered and the control point policy is involved in. A decomposition method is utilized, meanwhile, an idea of purchasing extra capacity is introduced to the hybrid system and a so-called blockage hedging point is obtained. Then a near-optimal method is presented to determine the real-time optimal buffer sizes if the maximum of the production rate of machines vary. Numerical results not only testify the method but also show us its application.

A Control Problem of Failure Prone Systems Considering Extra Capacity

The failure prone systems with the fixed maximum production rate, however uncertain demand levels, are examined. Each demand level may be larger or smaller than the maximum production rate of the systems. An optimal double hedging point (level) policy is proposed: a positive hedging point (inventory target) and a negative hedging point (backlog level below which extra capacity should be purchased). The analytical expressions of the positive and negative hedging points are obtained by a decomposition method. Meanwhile, a more straightforward necessary condition and a necessary and sufficient condition for the optimality of the just-in-time (JIT) control policy are presented respectively. A simple numerical comparison testifies and specifies our method