
This paper presents a designing methodology of repetitive group sampling plan to ensure the median lifetime of the products based on a new-compounded model called complementary exponentiated bell Touchard exponential distribution. A non-linear optimization model is constructed to determine the optimal sampling plan parameters, minimizing the average sample number while maintaining the desired quality assurance levels. This study aims to focus on the advantages of the proposed plan by comparing the performance of the OC curve and average sample number with various exponential families of lifetime models. To elucidate the performance of the proposed plan in industrial scenarios, real-life applications are presented. The results revealed that it may provide a better fit than other related lifetime models. The effect of misspecification of shape parameters is investigated and a sensitivity analysis is performed to understand how much the proposed plan is impacted. Furthermore, to reduce the overall inspection cost of the product, the economic designing of the proposed plan is meticulously analyzed under a complementary exponentiated bell Touchard exponential lifetime model.
The primary aim of this study is to propose a three-stage sequential sampling procedure for estimating the mean of a two-parameter exponential distribution under a bounded risk per unit cost constraint. The mean parameter theta of a two-parameter exponential distribution, is defined as alpha+beta where alpha and beta are the location and scale parameters, respectively. The constraint here is in such a way that the risk per unit cost is bounded below from a preassigned value omega. The expressions for the associated properties are derived according to the final sample size N. The simulation study is also conducted to verify the results of the crucial properties of the three-stage sequential procedure based on the risk per unit cost constraint and to compare the results of three-stage procedure with two-stage procedure due to Mahmoudi, Nemati, and Khalifeh. In the end, The performance of the suggested procedure is evaluated with the help of a real data set. Based on the obtained results, we conclude that the proposed three-stage procedure performs better and precisely.
Detecting structural changes in evolving networks is an important problem in sequential statistical inference. In many modern applications, interactions arise among groups rather than pairs, so hypergraphs can capture such higher-order relationships more accurately than ordinary graphs. In this paper, we study sequential detection of abrupt structural changes in dynamic hypergraphs under the hypergraph beta -model. We focus on a stylized setting in which the network is initially homogeneous and a denser sub-hypergraph emerges after a single unknown change-point. To address this problem, we develop a cumulative sum (CUSUM) procedure for the case where both pre-change and post-change parameters are known, and a window-limited CUSUM (WL-CUSUM) procedure for the more realistic case where the post-change parameter is unknown and must be estimated online. Both procedures are constructed from degree-sequence observations, which are sufficient statistics under the hypergraph beta -model and are especially attractive when the full hypergraph structure is difficult to observe or use in practice. We establish asymptotic properties of the proposed procedures and study the choice of the window size in the unknown-parameter setting. Numerical experiments and a real-data illustration demonstrate the performance and practical relevance of the proposed methods.
The current truncated sequential test theory, which uses the Expected Test Time (ETT) as the optimality criterion, cannot design customized test plans for different product cost structures. To address this issue, this paper proposes an Optimal Truncated Sequential Test (OTST) for the Mean Time Between Failures (MTBF, theta ) of exponential distributions, using a Comprehensive Expected Testing Cost (CETC) criterion instead of the traditional ETT. By analyzing the cost structure of truncated sequential tests, a cost model that incorporates fixed costs, time-dependent costs, per-item costs, and preparation costs related to the truncated sample size is constructed. Then, the theory of OTST based on CETC is established. Combining dynamic programming, the Nelder-Mead simplex method, and truncation sample size optimization, a solution method for the OTST is developed. Furthermore, the impact of the cost structure on the OTST is investigated. Results show that the proposed OTST adapts to different testing cost structures to meet product-specific testing needs. Through comparison with the test plans given in the IEC 61124 and GOST R 27.402 standards, as the Type I and Type II error probabilities are strictly controlled, the proposed OTST can save CETC by up to 16.72%. Moreover, a case study on aircraft black box reliability testing is presented to demonstrate the application process of the proposed OTST. The results indicate that, compared with the test plan of IEC 61124, the proposed method reduces CETC by approximately 3.04% (8,211 USD).