The mean time between failures (MTBF) and mean time to repair (MTTR) of manufacturing equipment (e.g., machines) are used in every quantitative method for production systems performance analysis, continuous improvement, and design. Unfortunately, the literature offers no methods for evaluating the smallest number of up- and downtime measurements necessary and sufficient to calculate reliable estimates of these equipment characteristics. This article is intended to provide such a method. The approach is based on introducing the notion of$(\alpha, \beta)$-precise estimates, where$\alpha $characterizes the estimate’s accuracy and$\beta $its probability. Using this notion, this article evaluates the critical number,$n^{*} (\alpha,\beta)$, of up- and downtime measurements necessary and sufficient to calculate$(\alpha, \beta)$-precise estimates of MTBF and MTTR. In addition, this article derives a probabilistic upper bound of the observation time required to collect$n^{*} (\alpha,\beta)$measurements.Note to Practitioners—To evaluate and predict production systems behavior, managers of manufacturing operations need to know equipment reliability characteristics. Quantifying the equipment status by MTBF and MTTR, this article provides answers to the following questions: Q1: How many measurements of machines up- and downtime are required to obtain reliable estimates of MTBF and MTTR? Q2: How long the observation period must be to collect the desired number of measurements? The answer to Q1 is provided by a rule (formula), which is based on the desired estimate accuracy (characterized by$\alpha $) and its likelihood (quantified by$\beta $). The answer to Q2 consists of selecting a small number of initial measurements, which can be used to calculate an upper bound of the total observation time.

Resilient monitoring systems (RMSs) are sensor networks that degrade gracefully under cyber-attacks on their sensors. The recently developed RMSs, while being effective in the attacked sensors identification and isolation, exhibited a drawback in their operation-an exponentially increasing assessment time as a function of the number of sensors in the network. To combat this curse of dimensionality, a decomposition approach has been proposed, which led to a dramatic reduction of the assessment time, irrespective of the sensor network dimensionality. However, information losses and, thus, reductions in the level of resiliency due to the decomposition, have not been investigated. This paper is intended to carry out such an investigation. Specifically, it derives conditions for lossless decomposition in terms of the Renyi-2 entropy. The development is based on the analysis of matrices, which characterize coupling of process variables and on a monotonicity property of the Dempster-Shafer combination rule on a class of functions, which arise within the RMS operation.

Production Lead Time \mbi (LT) is the average time a part spends in the system, being processed or waiting for processing. In systems with unlimited buffers, \mbi LT may be orders of magnitude larger than the total processing time, leading to serious economic and quality problems. At present, no systematic analytical methods for evaluation, analysis, and control of \mbi LT in systems with machines having up- and downtime characterized by continuous random variables are available. This paper is intended to develop such methods. Specifically, we address synchronous serial lines with exponential machines and derive formulas for \mbi LT as a function of machine parameters and raw material release rate. Using these formulas, we develop methods for open- and closed-loop raw material release, which result in the desired \mbi LT. For asynchronous exponential lines, we provide an upper bound on LT. For non-exponential lines (e.g., Weibull, gamma, and log-normal), we offer an empirical formula for \mbi LT as an affine function of the coefficient of variation. The results reported in this paper enable a new paradigm for production systems management, namely: manage a production system so that the desired \mbi LT is ensured, while the throughput is maximized. This paper analyzes production lead time in manufacturing systems with hardware-unlimited buffers. The main practical insights obtained are as follows. . The lead time as a function of the raw material release rate has a “knee-type” behavior. . Releasing raw material beyond the knee results in no noticeable increase of the throughput, but in unlimited increase of the lead time. . To maintain the lead time close to the knee (and, thus, maximize the throughput), a simple control law may be used: release raw material on, say, hourly or shift-type basis, if the total work-in-process is below a certain threshold (provided in the paper) and do not release otherwise.

Cellular lines are production systems consisting of cells comprised of machines performing similar operations. These systems are notorious for having excessively long lead time (LT) that are often an order of magnitude longer than the total processing time by all machines in the system. The main goal of this paper is to carry out an analytical investigation of this phenomenon and offer methods for its alleviation. To accomplish this, we develop a technique for performance evaluation of cellular lines with asymptotically reliable Bernoulli machines and use it for analysis and control of LT.

Resilient monitoring systems, considered in this paper, are sensor networks that degrade gracefully under malicious attacks on their sensors, causing them to project misleading information. The goal of this paper is to design, analyze, and evaluate the performance of a resilient monitoring system intended to monitor plant conditions (normal or anomalous). The architecture developed consists of four layers: data quality assessment, process variable assessment, plant condition assessment, and sensor network adaptation. Each of these layers is analyzed by either analytical or numerical tools. The performance of the overall system is evaluated using a simplified boiler/turbine plant. The measure of resiliency is quantified based on the Kullback-Leibler divergence and shown to be sufficiently high in all scenarios considered.

This paper describes the design and implementation of a continuous improvement project at an automotive paint shop. The development is based on a recently developed improvement methodology for production systems with quality-quantity coupled operations, whereby increasing the probability to complete a job during a cycle time leads to decreasing job quality. The implementation of the first stage of the project resulted in over 9% throughput improvement. The second stage predicts additional improvement of 9.5%.

The bottleneck (BN) of a production system is a machine with the strongest effect on the system's throughput. In this paper, a method for BN identification in serial lines with rework and Bernoulli machines is developed. The method can be applied using either calculated or measured data on blockages and starvations of the machines. For the case of calculated data, a technique for evaluating performance measures of Bernoulli lines with rework is developed. Along with these quantitative contributions, the paper provides three qualitative results. First, it shows that Bernoulli lines with rework do not observe the property of reversibility. Second, it demonstrates that downstream machines may have a larger effect on the throughput than upstream ones. Third, it demonstrates that BNs may be shifting not only because of changes in machine and buffer parameters but also due to changes in quality of parts produced.

In closed production lines, each part is placed on a carrier at the input of the first machine and is removed from the carrier at the output of the last machine. The first machine is starved if no carriers are available, and the last machine is blocked if the empty carrier buffer is full. The number of carriers in the system isSand the capacity of the empty carrier buffer isN0. Under the assumption that the machines obey the Bernoulli reliability model, this paper provides methods for determining if a pair (N0,S) impedes the open line performance and, if it does, develops techniques for improvement with respect toSandN0. In addition, bottlenecks in closed lines are discussed, and an approach to selecting the smallestN0andS, which result in no impediment, is described.

Lean buffering is the smallest buffer capacity, which is necessary and sufficient to ensure the desired production rate of a manufacturing system. Literature offers methods for designing lean buffering in production systems with identical machines. The current paper extends these methods to serial production lines with nonidentical machines, assuming that they obey the exponential reliability model. For two-machine lines, exact formulas for lean buffering are derived, while for longer lines estimates are obtained. These results can be useful for production line designers and production managers to maintain the required production rate with the smallest possible inventories. Note to Practitioners-In production systems with unreliable machines, operation with no-buffers (i.e., JIT) leads to low throughput. Very large buffers lead to high throughput but undesirable quality and economics properties. So, which level of buffering is good? This is the question addressed in this paper. The good level of buffering is addressed in terms of line efficiency, i.e., the fraction of the maximum of the throughput, which is acceptable for the system. For example, assume that the desired line efficiency is 0.9 (i.e., 90% of the maximum throughput is viewed as satisfactory). Under such an assumption, this paper offers methods for calculating the smallest (i.e., lean) level of buffering, which guarantees the desired throughput, provided that uptime and downtime of the machines are distributed exponentially with arbitrary mean time to failure and mean time to repair.

We prove that for any modulation, demodulation, and coding schemes and under general assumptions on the wireless channel model, the optimal power-efficient policy for transmit power assignment is necessarily of threshold nature. Although detailed features of this policy are quite complicated, it admits a simple and practical suboptimal version. Since both the optimal and suboptimal policies may have poor location-fairness characteristics, we introduce novel performance metrics (e.g., the average downtime and uptime), which quantify the short-term behavior of these policies, and develop analytical methods for their evaluation. Using these methods, we show that, although threshold policies may be extremely power efficient, users in poor locations relative to the base station may have unacceptable short-term performance characteristics. To alleviate this deficiency, we introduce an adaptive threshold policy, in which the threshold is adjusted to the user's location, and show that it is both relatively power efficient and location fair.

This paper develops an uplink transmission scheduling scheme, referred to as Ranking CDMA, which selects a subset of users for transmission at each time slot. We introduce long- and short-term metrics that characterize its performance, and devise analytical methods for evaluating these performance measures. It turns out that, although Ranking CDMA has excellent long-term power efficiency compared to traditional CDMA, it suffers from a lack of short-term fairness, which is quantified in our calculations. To alleviate this deficiency, we propose Adaptive Ranking CDMA and analytically show that it yields moderate power-efficiency improvement over traditional CDMA, while maintaining short-term fairness.

In this paper, we prove that the optimal power-efficient transmission policy, which ensures a desired throughput in a wireless network, is necessarily of threshold nature. Although detailed properties of this policy might be quite complicated, we show that it can be approximated by a suboptimal one, which is both simple and practical. This suboptimal policy may lead to substantial improvement in power-efficiency (compared to the widely used constant SNR policy), but at the expense of location-fairness. To alleviate this deficiency, we introduce an adaptive threshold policy and show that it is both relatively power-efficient and location-fair.

The problem of customer demand satisfaction in production systems with unreliable machines and finite finished goods buffers (FGB) is addressed. The measure of customer demand satisfaction is characterized by the probability to ship to the customer a required number of parts during a fixed time interval. This measure, referred to as the due-time performance (DTP), is often used to characterize the quality of a supplier in the automotive industry supply chain. In the paper, a method for evaluating DTP in serial and assembly lines is developed and the problem of selecting capacity of the FGB is discussed. The results obtained are illustrated by a case study at an automotive component plant.

The problem of satisfying customers demand in pull serial production lines is addressed. The level of demand satisfaction is quantified by the due-time performance (DTP), which is defined as the probability to ship to the customer a required number of parts during a fixed time interval. The definitions of DTP bottlenecks are introduced and a method for their identification is developed.

The bottleneck (BN) of a production line is a machine that impedes the system's performance in the strongest manner. It has been shown in previous work that the so-called downtime bottlenecks (DT-BNs) are of main industrial importance. In this paper, a method for DT-BN identification is developed in the framework of serial production lines with unreliable (Markovian) machines and finite buffers. The identification tool derived is based on the probabilities of machine blockages and starvations. To evaluate these probabilities, a novel aggregation technique is proposed. The results obtained are applied to the design of a production line at an automotive component plant.

A system-theoretic approach to bottlenecks in Markovian production lines is introduced and analyzed. The approach is based on the sensitivity of the system production rate to machines' reliability parameters. Using this approach, definitions of bottlenecks are introduced, methods for their identification are developed, and their implications for production automation and preventative maintenance are discussed.

Addresses a rate-based feedback approach to congestion control in packet switching networks where sources adjust their transmission rate in response to feedback information from the network nodes. Specifically, a controller structure and system architecture are introduced and the analysis of the resulting closed loop system is presented. Conditions for asymptotic stability are derived. A design technique for the controller gains is developed and an illustrative example is considered. The results show that, under appropriately selected control gains, a stable (nonoscillatory) operation of store-and-forward packet switching networks with feedback congestion control is possible.>

The theory developed by J.-T. Lim and S.M. Meerkov (see ibid., vol.COM-35, p.1278-88, Dec. 1987) is extended to channels with long propagation delays. The performance characteristics of Markovian access control protocols in satellite channels are evaluated. It is shown that the channel propagation delay may lead to improvements in systems behavior, i.e. to elimination of the saturation effects and to stabilization of the network in a steady state with relatively high performance characteristics.>

In this paper, a unified, analytical framework for analysis of Markovian access control protocols in collision channels with a finite number of buffered users is presented. The theory developed is based on: 1) an appropriate state-space modeling of a network, 2) the averaging theory for the slow-in-the-average Markov walks used for the local analysis, and 3) an asymptotic large deviations theory used for the global analysis. Specific new results obtained here are: 1) formulas for the steady-state buffer OCCupancy and the probability of packet rejection, 2) a relationship between the capacity of the users buffers and the steady-state throughput, time delay, and the probability of packet rejection, 3) an asymptotic formula for the saturation period, and 4) a new adaptive protocol, the adaptive TDMA, is constructed, and it is shown that it is superior to the adaptive ALOHA system.