Science AI Journal
Engineering · Research topic

Open research questions in Reliability and Maintenance Optimization

47 unresolved questions extracted from the limitations and future-work sections of 426 Reliability and Maintenance Optimization papers in our library. Each links back to the study that raised it.

What the literature leaves open

  • Environmental stressors in tropical regions. Limited resources in community-funded schools. The need for specialized technical expertise for restoration.

    Diagnostic framework and restoration protocols for visual display units: A reliability study of educational infrastructure in Edo State, Nigeria · 2026 · DOI

  • Safety-critical domains, such as power transmission and industrial production lines, require reliable systems. The k-out-of-n system needs to be optimized for reliability modeling and maintenance strategy. Intelligent maintenance approaches are needed.

    Reliability and Maintenance Optimization for $k$-out-of-$n$ Systems: A Systematic Review and Recent Advances in Theory and Practice · 2026 · DOI

  • Roof collapse in underground coal mines. The need to reduce production costs and mitigate negative environmental impacts. The complexity of the powered roof support system.

    RAM MEASURE OF HYDRAULIC POWERED ROOF SUPPORTS IN UNDERGROUND LONGWALL COAL MINES · 2026 · DOI

  • Cascading failure is a phenomenon in which an initial failure event propagates through functional connections and induces subsequent failures in other subsystems. Existing PageRank-based cascading failure models have limitations. There is a need for a fully data-driven approach to cascading failure analysis.

    Data-Driven Modeling of Cascading Failures via Bayesian Networks and PageRank Algorithm · 2026 · DOI

  • Limited test mileage and high reliability result in no-failure data during operational tests. The need to incorporate developmental test information into the reliability assessment. The lack of effective methods for reliability assessment under zero-failure data.

    A Hierarchical Bayesian Reliability Assessment Method for Vehicle Operational Tests under Zero-Failure Data · 2026 · DOI

  • Future developments in smart diagnostics, IoT-enabled monitoring, artificial intelligence, and digital twin technologies are expected to further improve the reliability and maintenance efficiency of automobile differential systems.

    Failure Analysis and Maintenance Assessment of Differential Systems in Automobile Power Transmission · 2026 · DOI

  • The need for a comprehensive assessment of failure mechanisms in automobile differential systems. The lack of evaluation of suitable maintenance strategies for improving reliability and service life.

    Failure Analysis and Maintenance Assessment of Differential Systems in Automobile Power Transmission · 2026 · DOI

  • Unpredictable component failures. Limited maintenance resources. Reactive maintenance practices.

    PROBABILISTIC ENHANCEMENT OF AN ISLAND NETWORK USING RELIABILITY-CENTERED MAINTENANCE · 2026 · DOI

  • This article has introduced a conceptual framework for embedding Bayesian Neural Network (BNN)–driven uncertainty quantification directly into maintenance‐logistics decisions. By reframing the classic Remaining Useful Life (RUL) question as a confidence‐based decision rule — “How confident are we that we can safely do nothing?” — the study bridges the long‑standing gap between predictive‑maintenance analytics and the day‑to‑day realities of spare‑parts planning and technician dispatching. 6.1 Key contributions • Unified decision logic – We formalise a single likelihood‑ratio threshold that transforms raw BNN predictive distributions into binary defer/act decisions and, in turn, into tangible logistics actions. ~ 171 ~ Published by Acta Logistica, www.actalogistica.eu Acta Logistica, Volume 13, Issue 1, Pages 162-173, 2026 ISSN 1339-5629 Applying Bayesian Neural Networks to optimize maintenance logistics Marc Hermans, Peter Tamas • Maintenance‑cost perspective – The approach explicitly links prediction uncertainty to cost elements (downtime, stock, overtime), enabling managers to reason in monetary terms rather than abstract accuracy metrics. • Generalisability – Because the framework is model‑agnostic apart from requiring predictive probabilities, it can incorporate future advances in probabilistic forecasting with minimal process changes. 6.2 Limitations Although the theoretical underpinnings are laid out in full, the framework has not yet been validated on real‑world or large‑scale synthetic datasets. The present work therefore stops short of providing empirical evidence of cost savings or service‑level improvements. Furthermore, computational considerations (latency, hardware footprint) are analysed only qualitatively at this stage. 6.3 Future work • Prototype implementation – We are developing a minimally viable BNN service layer and decision API to be tested on a live assembly line; initial latency targets are <200 ms per decision on industrial PCs. • Simulation‑based stress tests – A discrete‑event simulation seeded with actual MTBF/MTTR distributions will be used to quantify inventory reduction and uptime gains across a year‑long horizon. • Sensitivity analysis of the confidence threshold (β) – ROC‑style curves will help translate risk appetite into numerical β values for different asset classes. • Field deployment & benchmarking – Partner plants in the automotive and FMCG sectors have agreed to run A/B pilots comparing standard rule‑based scheduling with the BNN‑enabled logic. • The synthetic results in Figures 7 to 9 provide preliminary evidence of the BNN’s potential to reduce maintenance cost, spare parts inventory, and dispatch overhead. These findings will guide the simulation-based stress tests described in our future work, validating both the economic and logistical performance under realistic failure and demand scenarios. 6.4 Closing remarks Predictive‑maintenance research has advanced rapidly, yet many factories still rely on static service intervals because the operational impact of uncertainty remains opaque. By converting probabilistic predictions into economically meaningful actions, the proposed framework offers a pragmatic path to break that impasse [19]. The forthcoming implementation and validation phases will determine the extent to which the anticipated inventory and downtime benefits—often quoted but rarely measured—can be realised in practice.

    Applying Bayesian Neural Networks to optimize maintenance logistics · 2026 · DOI

  • The enhanced SARSOP algorithm achieved one order of magnitude reduction in α-vectors compared to PBVI, but the impact of aggressive pruning on solution quality under different convergence thresholds (other than ε=0.01) has not been explored for inspection-maintenance POMDPs.

    An optimal bi-level inspection and maintenance policy for a multi-component system: An enhanced successively approximated point-based value-iteration algorithm · 2026 · DOI

  • The comparison with alternative IM policies (TICBM, PICBM, PITBM, TBMA) is limited to a single initial condition (brand-new system state). The sensitivity of the BLIM policy's performance advantage across different initial belief states and varying inspection/maintenance cost ratios remains unexamined.

    An optimal bi-level inspection and maintenance policy for a multi-component system: An enhanced successively approximated point-based value-iteration algorithm · 2026 · DOI

  • Originality/value This study makes a distinct contribution by filling a significant gap in the literature using a Bayesian network (BN) to model the failure behavior of air jet looms machine using integrated quantitative failure/repair data and qualitative expert insights on human factors.

    Stochastic-based failure modeling of air jet loom machine · 2026 · DOI

  • To apply the proposed model to other complex data-driven systems. To compare the proposed model with other reliability models. To validate the proposed model using larger datasets.

    Smart Reliability Estimation Via ANN–ABC Optimization: A Novel Approach to Inverse Weibull Process Under NHPP · 2026 · DOI

  • Fig. 2: Comparative Estimation of the Cumulative Failure Function for 194 Men Aged 40 and Above Using MLE and ANN-ABC Methods Under the Inverse Weibull Process Model (DS1). A hybrid estimation of an Artificial Bee Colony (ABC) algorithm was compared as a possible combination with Artificial Neural Network (ANN) as an efficient framework to estimate the parameters in the Inverse Weibull Process (IWP) under a Non-Homogeneous Poisson Process (NHPP) distribution. Conducted with simulated as well as real world clinical failure data, the ANN-ABC method showed better accuracy when compared to the classical Maximum Likelihood Estimation (MLE) technique in terms of input data size; it is found to be more accurate with moderate to large samples. Its efficiency and robustness were supported by lower values of Root Mean Squared Error (RMSE) and Bayesian Information Criterion (BIC) and confirmed that it was able to capture time-dependent failure behavior. The results attest to the possibility of sophisticated algorithms in boosting reliability modeling of complex data-driven systems like predictive maintenance in cyber-physical systems, health informatics and IoT-based surveillance systems. The hybrid ANN-ABC model can be described as a scalable and adaptive mechanism of estimation that can be applied reliable performance of the traditional estimation approaches. The priority of future work will be to combine the Bayesian inference, imperfect repair modeling, and generalize it to the streaming or real-time data accommodating systems that are dynamic. Also, the next step in this project can be developing adversarial robustness and uncertainty quantification approaches to AI-driven reliability modeling supporting intelligent, secure, and trustworthy infrastructure. real-world deployment without in Fig. 3: Comparative Estimation of the Cumulative Failure Function for 105 women Aged 40 and Above Using MLE and ANN-ABC Methods Under the Inverse Weibull Process Model (DS2). The estimates of cumulative failure functions of the male © 2026 NSP Natural Sciences Publishing Cor.

    Smart Reliability Estimation Via ANN–ABC Optimization: A Novel Approach to Inverse Weibull Process Under NHPP · 2026 · DOI

  • Most existing models perform failure modes and RUL prediction independently, ignoring the inherent relationship between these two tasks. The lack of a unified approach to jointly model multi-sensor time-series data and failure event data.

    Bayesian Joint Model of Multi-Sensor and Failure Event Data for Multi-Mode Failure Prediction · 2026 · DOI

  • Further study on the hardware diversity of larger urban universities in Nigeria. Investigation of the applicability of the diagnostic framework to other types of computer hardware. Development of more advanced diagnostic tools and techniques.

    Diagnostic framework and restoration protocols for visual display units: A reliability study of educational infrastructure in Edo State, Nigeria · 2026 · DOI

  • Future research should explore integrating graph neural networks with deep neural networks. The paper suggests exploring the application of the proposed approach to various domains.

    Enhancing a multilayer perceptron model for multi-state network reliability evaluation via an arc-wise architecture · 2026 · DOI

  • The paper identifies the need for a more efficient and accurate method for MSN reliability estimation. The existing methods have limitations, such as requiring MPs or d-MPs in advance.

    Enhancing a multilayer perceptron model for multi-state network reliability evaluation via an arc-wise architecture · 2026 · DOI

  • There is a need to optimize replacement cycles of mechanical components in large-scale electric excavators. Current methods may not be effective in evaluating reliability and optimizing replacement cycles.

    Metodologia de aumento de vida útil: aplicação em ativos de escavação de grande porte · 2026 · DOI

  • The need to assess the reliability and maintainability of powered support equipment to maintain continuity in mining operations and reduce costs. The lack of efficient maintenance schedules for powered support systems.

    RAM MEASURE OF HYDRAULIC POWERED ROOF SUPPORTS IN UNDERGROUND LONGWALL COAL MINES · 2026 · DOI

  • This study proposed a data-driven BN–PageRank framework for cascading failure modeling in complex repairable systems. The framework addresses a key limitation of existing PageRank-based cascading failure analysis, namely its reliance on expert-defined propagation networks and manually assigned damping factors. By learning directed failure propagation structures from field failure records through score-based Bayesian Network structure learning and estimating the damping factor from observed transition patterns via maximum likelihood estimation, the proposed approach enables a more objective and reproducible implementation of cascading failure analysis. The proposed framework was applied to warranty claim records from CNC machining centers consisting of eleven functional subsystems. The observed failure records were transformed into a binary panel representation that captures both current failure occurrence and prior subsystem failure history. The learned Bayesian Network provided a data-supported representation of candidate failure propagation pathways, and the estimated damping factor quantified the extent to which observed failure transitions were attributable to network-based propagation. The resulting PageRank scores were then incorporated into the reliability decomposition model inherent subsystem failure behavior from propagation-induced failure contributions. to distinguish findings suggest The empirical results demonstrate that the proposed framework provides interpretable subsystem-level propagation information, improving failure count prediction accuracy compared with a conventional reliability model. In particular, the proposed BN– PageRank model substantially reduced prediction errors during the validation period relative to the conventional Weibull model.

    Data-Driven Modeling of Cascading Failures via Bayesian Networks and PageRank Algorithm · 2026 · DOI

  • The study uses a simulated dataset, which may not reflect real-world scenarios. The sample size is limited to 1000, which may not be representative of larger populations.

    Fuzzy semi-Markovian stochastic model for single-unit system with repairman arrival delay under Lindley lifetime distribution using bell-shaped membership function · 2026 · DOI

  • Future research can focus on extending the proposed model to more complex systems. The study's results can be used as a basis for further research on the impact of fuzziness and repairman arrival delay on system performance.

    Fuzzy semi-Markovian stochastic model for single-unit system with repairman arrival delay under Lindley lifetime distribution using bell-shaped membership function · 2026 · DOI

  • The limitations of existing degradation-centric models overlook the random failure characteristics of electronic components. The storage reliability management of advanced missile systems is a key challenge.

    A Study on an ASRP Based Approach for Guided Missiles Considering Electronic Component Failure Characteristics · 2026 · DOI

  • To apply the proposed method to other domains with similar challenges. To develop more advanced methods that can incorporate additional sources of information. To investigate the use of other lifetime models and prior distributions.

    A Hierarchical Bayesian Reliability Assessment Method for Vehicle Operational Tests under Zero-Failure Data · 2026 · DOI

Most-cited papers in Reliability and Maintenance Optimization

Most recent work

Find a gap in your own Reliability and Maintenance Optimization sub-topic

This page shows what the Reliability and Maintenance Optimization literature already flags as unresolved. To narrow it to your specific question, run the guided finder — it searches the gap library on demand and checks candidates against 250M+ OpenAlex works.

Open the Research Gap Finder →

Related topics in Engineering

Traffic and Road Safety
237 open questions
Concrete and Cement Materials Research
196 open questions
Digital Transformation in Industry
190 open questions
Design Education and Practice
167 open questions
Machine Fault Diagnosis Techniques
165 open questions
Smart Cities and Technologies
152 open questions
Spatial Cognition and Navigation
149 open questions
BIM and Construction Integration
145 open questions
47 open questions have been extracted from the limitations and future-work passages of 426 Reliability and Maintenance Optimization papers in our library. Each one below links back to the study that raised it, so you can read the original claim in context.

Tools for your next paper

Gaps & Literature SearchFind open questions in any field — 120,000+ open gaps mined from our own library.Pre-CheckIs your paper ready? A Tier 1–5 acceptance probability in seconds.Journal FinderA ranked shortlist of target journals from a 17,500-venue index.Duplicate Publication CheckerCheck prior-publication & salami-slicing risk across 8 sources.Citation GeneratorFreeA DOI or title → APA, MLA, Vancouver, Chicago, BibTeX, RIS. No signup.Graphical Abstract MakerFreeTurn your findings into a colorblind-safe graphical abstract.

Compare the category — Honest roundups of the AI research tools, ours listed alongside the alternatives.

Command palette

Jump anywhere, run any action.