Humanitarian crises are escalating globally, with disaster-affected populations increasing by 23% annually. Yet traditional single-vehicle logistics systems fail catastrophically when infrastructure is compromised, creating life-threatening delays in critical supply delivery. While emerging truck-drone integrated systems show promise, existing research remains fragmented across disciplines, lacking systematic analysis of risk-aware optimization approaches and their real-world applicability. This study addresses two primary research objectives: (1) to develop a comprehensive taxonomy of mathematical models and solution methodologies for truck-drone humanitarian logistics, and (2) to quantify performance improvements and identify implementation barriers. A systematic review of 132 peer-reviewed papers from Scopus (2020–2026) was conducted using PRISMA methodology, complemented by bibliometric analysis of citation networks. The analysis reveals that hybrid solution approaches dominate the field (67 studies), with emergency response applications showing the strongest research focus (36 studies), though only about 40% report validation beyond purely simulated experiments. This review identifies three main contributions: a structured taxonomy of 15 distinct problem formulations and 8 solution categories, a quantitative benchmarking framework revealing performance trade-offs, and identification of critical research gaps in uncertainty modelling and stakeholder coordination. The results inform policy frameworks for regulatory approval while offering practitioners evidence-based counsel for decisions about technology adoption. This analysis lays the groundwork for more robust humanitarian logistics networks by advancing both theoretical knowledge and actual application of risk-aware truck-drone systems.
Given the substantial demand for hazardous material transportation and the potential risks to the public, transportation optimization research is crucial for reducing accidents involving the transport of hazardous materials. We investigate a multi-depot location-routing problem considering the risk impact (MDLRP-RI) for the design of a novel hazardous material transportation system. It is formulated as a mathematical model to minimize the total travel time and transportation risk. A hybrid tabu search algorithm is proposed based on the spatial-risk coupling clustering method (HTS-CM) with a specifically designed initial solution, efficient local search operators, diversification strategies to solve this MDLRP-RI optimization problem. Various experimental instances are conducted to demonstrate the effectiveness of the HTS-CM algorithm. The results of multi-depot location routing and the marginal trade-off analysis of Pareto solutions are discussed through an empirical case study. The MDLRP-RI strategy can reduce the weighted objective function ranging from 6.7% to 25.4%. We also provide some managerial insights and identify the impact of key parameters on the results in the sensitivity analysis.
Smart manufacturing increasingly requires design teams to coordinate mechanical, behavioral, control, and performance dimensions in parallel. However, when design changes occur, their cascading effects often remain implicit, embedded in undocumented assumptions, and resolved only through repeated cross-disciplinary negotiation. The fundamental challenge lies in the lack of formally defined dependency relationships across design dimensions, which makes coordination dependent on human judgment and costly iterations. This study proposes the Quad-Play Reference Model (QPRM), which establishes a formal cross-dimensional design reference integrating the Structure, Behavior, Control, and Performance dimensions. Through structured association logic that characterizes both cross-dimensional and hierarchical relationships, the QPRM makes design content dependencies explicit and traceable, thereby supporting coordinated design and providing a structured basis for identifying the scope of design changes. The QPRM instantiation architecture further specifies how design content is constructed and linked across dimensions, creating the conditions for meaningful GenAI assistance in design realization. A bi-level agent framework is developed to operationalize this process by translating design intent into platform-executable actions through a Logic-Syntax Decoupling architecture. Validation through a chip System-Level Test equipment case study demonstrates that the QPRM substantially reduces coordination complexity and iteration risk. Complementary experiments on production-line cases at multiple granularities further confirm the practical feasibility of GenAI-assisted QPRM instantiation.
