Expected Failure Rate Research Articles

A note on infinite-armed Bernoulli bandit problems with generalized beta prior distributions

A bandit problem with infinitely many Bernoulli arms is considered. The parameters of Bernoulli arms are independent and identically distributed random variables from a generalized beta distributionG3B(a, b, λ) witha, b>0 and 0<λ<2. Under the generalized beta prior distributions, we first derive the asymptotic expected failure rates ofk-failure strategies, and then obtain a lower bound for the expected failure rate over all strategies investigated in Berry et al. (1997). The asymptotic expected failure rates for the other three strategies studied in Berry et al. (1997) are also included. Numerical estimations for a variety of generalized beta prior distributions are presented to illustrate the performances of these strategies.

Results and lessons learned from a small medical abortion clinical study in Turkey

This article presents results from a small, introductory clinical study offering a simplified regimen medical abortion to women in Turkey. A total of 208 women from five sites were recruited from July 2000 through March 2001. All eligible consenting women were given 200 mg oral mifepristone at the clinic followed by 400 μg oral misoprostol 2 days later either at home or at the clinic. The overall success rate of 84.1% is lower than in previous studies of this regimen, and a surprisingly high proportion of women (9.1%) were diagnosed with incomplete abortion. In spite of the higher than expected failure rate, women expressed a high degree of satisfaction. Several challenges were faced during this study, providing valuable insights about ways to introduce medical abortion into new settings with the best chances for success.

Cumulative sum techniques for assessing surgical results

The report by Novick and colleagues [1], in this issue of The Annals of Thoracic Surgery uses a cumulative sum (CUSUM) technique to assess the learning curve in telerobotic surgery. Novick and associates [2] have previously used CUSUMs to describe the learning curve of an academic surgeon, the change from on-pump to off-pump coronary bypass surgery [3], and the learning curve for off-pump surgery [4]. From these studies, they found that CUSUM provided “ . . . a more sensitive indicator of a cluster of surgical failures than standard statistical techniques” [1].

Probabilistic evaluation of lightning-related failure rate of power system apparatus

Traditionally, insulation coordination of power system apparatus relies on deterministic approaches where some conservative voltage stresses are defined and the insulation levels are then adjusted using protective ratios and margins. In the case of lightning stresses, this could easily lead to overinsulation or overdesign and subsequently to an increase in the cost of equipment. On the other hand, insufficient insulation could lead to power system failures with disastrous technical and economic consequences. In this paper, we present a formalized method for predicting the lightning-related failure rate of power system equipment. In the probabilistic approach to insulation coordination of power system apparatus we quantify the expected risk of failure under lightning stresses to enable an economically optimized design. To illustrate the method, the expected failure rate of surge arresters applied to unshielded 230-kV transmission line has been calculated.

Randomized trial of oral versus vaginal misoprostol 2 days after mifepristone 200 mg for abortion up to 63 days of pregnancy

This prospective, open-label, randomized trial of healthy adult women up to 9 weeks pregnant compared mifepristone 200 mg followed 2 days later with misoprostol 400 μg orally versus misoprostol 800 μg vaginally. The study was interrupted after the oral misoprostol group experienced a higher than expected failure rate. This treatment was discontinued and another substituted consisting of oral misoprostol 800 μg divided into two doses two hours apart. Women returned for a follow-up visit from Day 4 to 8. All women with a continuing pregnancy received a repeat dose of misoprostol vaginally and returned before Day 15. The primary outcome measure was a complete medical abortion without surgical intervention at the first visit. Of the 1045 women enrolled, 1011 had complete data: Group 1 (220) used oral misoprostol 400 μg, Group 2 (269) used oral misoprostol 800 μg, and Group 3 (522) used vaginal misoprostol 800 μg. At first follow-up visit, the primary outcome, that is, a complete abortion, was 84% for Group 1, 92% for Group 2, and 96% for Group 3, p < 0.001. After a second dose of vaginal misoprostol in women with on-going pregnancies at their first follow-up visit, the complete abortion rates were 91%, 95%, and 98%, respectively, p < 0.001. There were minimal differences in side effects, onset of bleeding and overall acceptability in the three groups. Mifepristone 200 mg followed by vaginal misoprostol 2 days later was more effective at inducing an abortion up to 9 weeks of pregnancy than the same dose of mifepristone followed by oral misoprostol.

Predicting vegetation-related failure rates for overhead distribution feeders

Faults on the electric power distribution system are responsible for a large portion of the interruptions that a customer will experience. To maintain a high level of system reliability, vegetation maintenance is often required. Analytical prediction of the effects of vegetation maintenance on distribution system reliability requires a model of the expected failure rate of line sections that includes the effects of vegetation. Vegetation-related failures are more likely to occur as the vegetation near the overhead power lines grows, increasing the line-section failure rate. Due to difficulties in using existing growth models, this paper proposes to use a direct model for failure-rate prediction based on factors that affect vegetation growth. Four models are considered: linear regression, exponential regression, linear multivariable regression, and an artificial neural network (ANN). The models are tested with historical vegetation growth parameter data and feeder failure rates. Results are compared and the features of each model are discussed.

Predicting Vegetation Related Failure Rates for Overhead Distribution Feeders

Faults on the electric power distribution system are responsible for a large portion of the interruptions that a customer will experience. To maintain a high level of system reliability, vegetation maintenance is often required. Analytical prediction of the effects of vegetation maintenance on distribution system reliability requires a model of the expected failure rate of line sections that includes the effects of vegetation. Vegetation-related failures are more likely to occur as the vegetation near the overhead power lines grows, increasing the line-section failure rate. Due to difficulties in using existing growth models, this paper proposes to use a direct model for failure-rate prediction based on factors that affect vegetation growth. Four models are considered: linear regression, exponential regression, linear multivariable regression, and an artificial neural network (ANN). The models are tested with historical vegetation growth parameter data and feeder failure rates. Results are compared and the features of each model are discussed.

Consolidating Maintenance Spares

The inventory of spare parts that a firm holds depends on the number of working parts and age of the equipment to be serviced, the expected failure rate associated with each working part, and the acceptable level of service. We model the problem of consolidation of spare parts to reduce overall inventory as an integer program with a nonlinear objective function. A linear reformulation of this model is obtained that helps solve some practical instances. A more compact implicit formulation is developed and solved using a specialized branch-and-price technique. We also demonstrate how this specialized branch-and-price technique is modified to devise a very effective heuristic procedure with a prespecifiable guarantee of quality of solution produced. This provides a practical and efficient methodology for maintenance spare consolidation.

Calculating the expected failure rate of complex equipment subject to hazardous repair

In hazardous repair of a repairable system, there is a strong probability of further failure during a period immediately following the repair. To model failures of such systems under failure-based maintenance, a class of point process models is proposed in which the failure intensity is a function only of equipment age and of time since most recent repair. Cost modelling requires the computation of the expected number of failures by age t as a function of equipment age. This expected number is shown to satisfy a Volterra (integral) equation of the second kind that generalises the renewal equation. A solution algorithm is given. A simple (four-parameter) model from this class is introduced, and the methodology is exemplified using a database of failures of non-maintained medical equipment.

Elevation of the tibial tubercle for patellofemoral pain syndrome. An 8- to 15-year follow-up.

One hundred patients were operated on by the Maquet procedure for chondromalacia patellae. All of them were first re-examined after a mean follow-up of 4 years, and 65 of them re-examined after a mean follow-up of 11 years (range, 8-15 years). The pain score improved significantly after the operation and remained unchanged with longer follow-up. The success rate was only 62% at both follow-ups. Outerbridge grade IV chondral lesions at the time of surgery were associated with a significant improvement of the pain score at the 4-year follow-up and a success rate of 69%. The Maquet procedure should only be proposed for chronic retropatellar pain with grade IV chondral lesions, after conservative treatment has proven unsuccessful, as the expected failure rate is about 30%.

Oral contraception noncompliance: the extent of the problem.

Compliance has been defined as the extent to which a patient's behavior coincides with the clinical prescription. The lowest expected failure rate of oral contraceptive (OC) use has been suggested to be 0.1%; however, in typical users, the failure rate varies by age, race, and marital status. In some populations in the United States (eg, unmarried black adolescents), the failure rate is 18%; rates that factor in the occurrence of abortions suggest an overall failure rate of 6.2% during the first year of use. If the annual failure rate of 6.2% is assumed, long-term failure rates rapidly approach 25% to 50% (ie, the proportion of users who will experience an unintended pregnancy) over 10 years of use. The gap between the lowest expected failure rate and the failure rates seen in typical users is related in large part to problems of compliance. Compliance with OC use includes both correct daily use and continuing use. Factors that contribute to OC discontinuation include the experience of side effects, fears and misinformation about OC side effects fueled by negative media reports, and intermittent use. There are relatively few data regarding the issue of missed pills and correct use, but studies suggest that adolescents may miss an average of 3 pills a month, and at least 20% to 30% of individuals miss a pill every month. Patients may have difficulty in the transition from one packet of pills to the next, and missed pills that extend the hormone-free interval may contribute to the failure rate.(ABSTRACT TRUNCATED AT 250 WORDS)

Minimum Variance Unbiased Estimation of Software Reliability

As the formal methods of proving correctness of a computer program are still very inadequate, in practice when a new piece of software is developed and all obvious errors are removed, it is tested with different (random) inputs in order to detect the remaining errors and assess its quality. We suppose that whenever the program fails the error causing the failure can be detected and removed correctly. Thus, the quality of the software increases as testing goes on. In this paper, we consider two different models and present the minimum variance unbiased estimators of the expected failure rate of the revised software at any time of testing t, based on the data generated up to that point.

A Method for Predicting Results of Humidity Tests

The paper outlines a method for predicting the results of humidity tests applied to electronic equipments. It is based on a theory of vapor entry into parts to cause failure. Depending on the packaging, the entry mechanism and the test conditions, the rate of vapor entry can be calculated and a severity factor can be estimated for different test levels. If a base level is set at about 25°C (77°F) and 50% RH, a base failure rate can be derived for each part from the dormant or storage failure rates, which are well established. The expected failure rate under the test is derived from the base failure rate and the acceleration factor for each part in an equipment or unit to calculate a total failure rate. With the total failure rate and the test exposure time, the basic reliability equation can be used to calculate the probability of passing or failing the test. The method provides an estimate of the probability similar in principle and with an accuracy comparable to the reliability prediction methods of MIL-HDBKD-217. Examples of prediction and references are given.

Stress test development on a processor system

AbstractThermal cycling and vibration tests were carried out on an intermediate processor family of machines to investigate the benefits of controlled stressing as a means of screening manufacturing defects, and to provide data for return on investment calculations. Major items to be considered were(a) low cost of test programme(b) complete unit to be stressed (complex interactions between subassemblies)(c) relatively low expected failure rate(d) limited field reporting.Stress techniques used were mild thermal cycling and pseudo‐random vibration.Analysis of field data from the stressed population showed an overall reliability improvement of 24 per cent during the first 30 days of customer operation.

Estimate of the probability of a lightning strike to the Galileo probe

Lightning strikes to aerospace vehicles occur mainly in or near clouds. As the Galileo entry probe will pass most of its operational life in the clouds of Jupiter, which is known to have lightning activity, the present study is concerned with the risk of a lightning strike to the probe. A strike to the probe could cause physical damage to the structure and/or damage to the electronic equipment aboard the probe. It is thought to be possible, for instance, that the instrument failures which occurred on all four Pioneer Venus entry probes at an altitude of 12 km were due to an external electric discharge. The probability of a lightning strike to the Galileo probe is evaluated. It is found that the estimate of a strike to the probe is only 0.001, which is about the same as the expected failure rate due to other design factors. In the case of entry probes to cloud-covered planets, a consideration of measures for protecting the vehicle and its payload from lightning appears to be appropriate.

Performance of United States citizens with foreign medical education on standardized medical examinations.

To enter the mainstream of American medicine, United States students who enroll in foreign medical schools must either transfer to a United States medical school (usually after passing Part I of the NMBE examination) or complete their studies abroad and then pass FLEX and usually ECFMG examinations. Half those applying for transfer fail Part I. Although virtually all who succeed in transferring subsequently pass Parts II and III, their scores are below the mean of the reference group, and their relative performance progressively deteriorates. Sixty per cent of USFMG's fail the medical portion of the ECFMG examination as compared to an expected failure rate for United States medical students of only 2 to 3 per cent. Thirty-eight per cent of USFMG's who take both ECFMG and FLEX examinations fail the latter. The mean weighted score for USFMG's on FLEX is 75.8, while that for United States graduates is 80.1 and for FMG's 74.7. These preliminary data suggest that United States students who go abroad, regardless of where they complete their studies, do not perform as well on standard tests after they return home as do those who have had their entire education in United States schools. This question needs more systematic study.

Sequential Anastomoses in Coronary Artery Grafting: Technical Aspects and Early and Late Angiographic Results

The present study concerns itself with the early and late results obtained with aortocoronary vein grafts containing more than 1 coronary anastomosis per graft. The surgical technique is described in detail and some of the hazards are outlined. It is apparent that the use of side-to-side anastomoses (SSAs) leads to a marked increase in blood flow in the proximal portion of the graft (average flow, 131 ml/min) and also increases the patency rate of the proximal anastomosis. Thus, 3 of 51 SSAs were obstructed on the early angiogram, and only 1 of the 28 SSAs restudied at one year had become occluded. Although the cumulative—early and late—patency rate (55/66, or 83.3%) of the distal end-to-side anastomoses (ESAs) was not significantly better than that usually seen with conventional vein grafts, it is believed that improvement in the technique and greater awareness of some of its pitfalls will further decrease the occlusion rate of distal ESAs. Most occlusions of the ESA appeared related to angulation, which tended to occur in the segment of graft between the proximal and the distal anastomoses. Appropriate modifications in the technique are described that should help to eliminate this difficulty. The use of SSAs also allows for grafting of small coronary arteries which, with the conventional single anastomosis technique, are not usually bypassed in view of the high expected failure rate. Thus, of the 9 SSAs performed to 1 mm arteries and studied on two occasions, 8 remained open early after operation and 7 were still patent a year later.