Establishment and Solution of a Multi-Stage Decision Model Based on Hypothesis Testing and Dynamic Programming Algorithm
Ziyang Liu, Yurui Hu, Yihan Deng · 2025
Read Establishment and Solution of a Multi-Stage Decision Model Based on Hypothesis Testing and Dynamic Programming Algorithm by Ziyang Liu, Yurui Hu, Yihan Deng (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-05807.txt.
Introduction
Ziyang Liu, Yurui Hu, and Yihan Deng build a step-by-step decision model that joins hypothesis testing with dynamic programming. The model tackles choices made in stages, where each step depends on what came before. It starts with a sampling inspection plan that uses random sampling without replacement to judge a defect rate while keeping both kinds of testing error under control.
From the testing theory, the authors derive the spread of the sample defect rate and the smallest sample size needed in two settings. They then set up the staged model with state changes and goals for each stage, and they work backward through it to find the best plan. Six optimal strategies come out under different conditions. The authors argue the model stays easy to interpret while handling complex staged choices, which suits quality control and similar applied problems.
Catalog record
| CREATOR | Ziyang Liu, Yurui Hu, Yihan Deng |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 22 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-05807
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Establishment and Solution of a Multi-Stage Decision Model Based on Hypothesis Testing and Dynamic Programming Algorithm Ziyang Liu, Yurui Hu, Yihan Deng Abstract This paper introduces a novel multi-stage decision-making model that integrates hypothesis testing and dynamic programming algorithms to address complex decision-making scenarios.Initially,we develop a sampling inspection scheme that controls for both Type I and Type II errors using a simple random sampling method without replacement,ensuring the randomness and representativeness of the sample while minimizing selection bias.Through the application of hypothesis testing theory,a hypothesis testing model concerning the defect rate is established,and formulas for the approximate distribution of the sample defect rate and the minimum sample size required under two different scenarios are derived. Subsequently,a multi-stage dynamic programming decision model is constructed.This involves defining the state transition functions and stage-specific objective functions,followed by obtaining six optimal decision strategies under various conditions through backward recursion.The results demonstrate the model’s potent capability for multi-stage decision-making and its high interpretability,offering significant advantages in practical applications. Keywords Multi-stage decision-making model, Hypothesis testing, Dynamic programming algorithms I.Introduction Decision-making in complex environments often requires a methodical approach to handle uncertainties and dynamically changing conditions. Traditional decision models frequently fall short when faced with multi-stage scenarios where decisions at one stage influence outcomes in subsequent stages. To address this gap, we propose a comprehensive model that combines hypothesis testing with dynamic programming to formulate and solve multi-stage decision problems. This approach allows for a more nuanced control of decision-making errors and adapts to the evolving nature of decision scenarios. Zhu Chenlu and colleagues [ 1 ] initially employed probabilistic interval hesitant fuzzy sets for depicting the decision-making system’s hesitancy, followed by the development of an optimization model leveraging the score function, deviation function, and information entropy of these fuzzy elements to extract probabilistic information. Subsequently, to enhance the computational efficiency, a cloud model was adopted, facilitating the transformation from PIVHFS to this cloud model through an established optimization framework. Liu Weiqiao and Zhu Jianjun [ 2 ] introduced a multi-stage decision-making approach that incorporates quantum guidance, utilizing the normal cloud model for expert relocation. Zhong Xiangyu and associates [ 3 ] developed a novel consensus model for large-group multi-attribute decision-making, asserting the equal significance of basic and ordinal consensus by integrating these aspects. Mao Xingyun and team [ 4 ] introduced a preprocessing strategy aimed at simplifying the computational demands of the latent group lasso (LGL) issue, which notably reduces unnecessary overlapping support groups while maintaining a constant count of essential support groups. Li Lubo and collaborators [ 5 ] tackled the Manufacturing Project Scheduling Problem with Preparation Time in Dynamic and Disruptive Environments (MPSPST-DIE), a scenario frequently arising in uncertain and spatially disturbed production settings, by crafting a multi-agent genetic planning super heuristic (HH-MGP) algorithm for its resolution. Hypothesis testing, a statistical method used to make inferences about populations based on sample data, provides a structured framework to control and assess the risk of incorrect decisions (Type I and Type II errors). By applying these principles, we ensure that the decision-making process is both evidence-based and statistically sound. The incorporation of dynamic programming, a mathematical optimization method, allows for the efficient solving of complex problems by breaking them down into simpler, interrelated stages. This method’s forward-looking nature and its ability to consider the entirety of a decision path make it ideally suited for multi-stage decision-making. The synergy between hypothesis testing and dynamic programming in our model facilitates the handling of decision-making scenarios with a higher degree of complexity and uncertainty. By systematically addressing the statistical and sequential aspects of decision-making, our model is capable of providing robust solutions that are both optimal and practical. II.Design of sampling test methods A.Hypothesis testing and approximate distribution Hypothesis testing Hypothesis testing is the process of statistically analyzing sample data to determine whether a hypothesis (usually about an overall parameter) is valid. In this problem, the supplier claims that the defective rate of spare parts will not exceed a certain nominal value, and hypothesis testing can be used to assess whether the sample data supports this claim by the supplier. In this problem, the following two hypotheses can be constructed: Original assumption H_{0}:p\leq p_{0} , the rate of defective products does not exceed the nominal value claimed by the supplier; Alternative hypothesis H_{1}:p\geq p_{0} , the rate of defective products exceeds the nominal value claimed by the supplier. Approximate distribution Let the total number of spare parts be , in which each spare part can only be qualified or non-qualified products, so we can specify the following overall unit index: Y_{i}=\begin{cases}1,&\text{If the spare parts are not qualified,}\\ 0,&\text{If the spare parts are qualified.}\end{cases} (1) From the above only 0, 1 two kinds of index value of the overall take a sample size of n simple random sample, let be the number of substandard spare parts in the sample, then the sample defective rate: p=\frac{a}{n} (2) Which is the sample mean, and thus the sample variance: s^{2}=\frac{1}{n-1}\cdot\sum_{i=1}^{n}\cdot\left(y_{i}-\bar{y}\right)^{2}=\frac{n}{n-1}\cdot p\cdot\left(1-p\right) (3) And the variance of the sample defective rate is: v\left(p\right)=\frac{1-f}{n}\cdot s^{2}=\frac{1-f}{n-1}\cdot p\cdot\left(1-p\right) (4) Since no-return simple random sampling is used a=n\cdot p , obeys the hypergeometric distribution: a\sim\text{HG}(N,M,n) (5) The expectation of the hypergeometric distribution is: E(a)=n\cdot\frac{M}{N} (6) And the variance of the hypergeometric distribution is: V(a)=n\cdot\frac{M}{N}\cdot\left(1-\frac{M}{N}\right)\cdot\frac{N-n}{N-1} (7) where M is the number of nonconforming parts in the population N . It is useful to note that P=\frac{M}{N} , and P is the actual defective rate. \frac{N-n}{N-1} is the correction factor due to non-return sampling, as distinguished from the binomial distribution. When N is large, \frac{N-n}{N-1}\approx 1 , and a approximately follows the binomial distribution: a\sim\text{B}(n,P) (8) When n is also large, a again approximately follows a normal distribution: a\sim N(nP,nP\left(1-P\right)) (9) In turn, the distribution of the sample defective rate p can be obtained from equation (2): p\sim N\left(P,\frac{P(1-P)}{n}\right) (10) B.Sampling and testing programs to control Type I errors In this paper, we wish to design a sampling scheme that can determine whether to reject the lot of spare parts at a 95% confidence level. In this case, the firm wants to control that the probability of the first type of error (wrongly rejecting the lot of spare parts when H_{0} is true) \alpha does not exceed 0.05. Test statistic The sample defect rate p obeys a normal distribution if the actual defect rate P=p_{0} : p\sim N\left(p_{0},\frac{p_{0}(1-p_{0})}{n}\right) (1) In order to check whether the defective rate provided by the supplier exceeds the nominal value p0 , we standardize the sample defective rate p to obtain the test statistic Z : Z=\frac{\hat{p}-p_{0}}{\sqrt{\frac{p_{0}(1-p_{0})}{n}}} (2) At this point, Z obeys the standard normal distribution N(0,1) . From the standard normal distribution table, the critical value for the 95% confidence level is: Z_{0.95}=1.645 (3) To derive the minimum sample size, we set the permissible absolute error limit to d , i.e., we want the difference between the defective rate in the sample and the nominal defective rate p_{0} to be no more than d . At this point, the Z_{0.95}=\frac{d}{\sqrt{\frac{p_{0}(1-p_{0})}{n}}} (4) The final formula to obtain the sample size is: n=\frac{p_{0}(1-p_{0})}{\left(\frac{d}{Z_{0.95}}\right)^{2}} (5) d is in the range of 0.02 to 0.09, the sample size at each step of 0.01, varies as shown in