Zhiyun Zou

The new energy vehicle (NEV) industry anchors China’s transition towards sustainable mobility, and the Pearl River Delta (PRD) hosts one of its densest clusters. Yet the spatial organisation of the innovation system underpinning it remains poorly understood, because actors, data and methods are heterogeneous. This study develops a hybrid computational intelligence framework linking kernel density estimation, spatial autocorrelation, social network analysis and temporal exponential random graph models (TERGMs) in one reproducible pipeline, applied to 4036 innovation entities, 1260 co-applied NEV patents yielding 274 cooperating pairs among 290 organisations, and 187 policy documents (2009–2025). The chain exhibits a dual-core, core-periphery agglomeration (global Moran’s I = 0.538, p = 0.001): a Shenzhen-centred mass extending into Dongguan and a smaller Guangzhou core, against 14 low-low peripheral units. Its cooperation network expands rapidly without consolidating: cooperating organisations grow from 30 to 290 while density falls from 0.071 to 0.007; it stays sparse and modular, anchored by universities and flagship firms. Geographic proximity and organisational scale raise, and technological distance lowers, the odds of cooperation, while same-tier alignment is not significant once proximity is controlled. Contrary to expectations, cities with richer non-subsidy portfolios show weaker tie formation (β = −0.59, p < 0.001), and strengthening over time, whereas subsidy intensity is only weakly positive (β = 0.16, p = 0.067). These associations, stable across five specifications, are observational, not causal; they imply that the sustainability of the PRD’s NEV innovation system rests less on further concentration than on making existing proximity usable across municipal boundaries.

The new energy vehicle (NEV) industry is a strategic pillar of China’s manufacturing transition, and the Pearl River Delta (PRD) hosts one of its most concentrated innovation clusters. Yet the spatial layout, network structure, and policy drivers of the PRD NEV innovation chain remain poorly understood because the relevant actors, data, and methods are highly heterogeneous. This study proposes a hybrid computational intelligence framework that integrates kernel density estimation, spatial autocorrelation, social network analysis, and temporal exponential random graph models (TERGMs) to analyse multimodal data on 4,036 innovation entities, 95 patent-based cooperation ties, and 187 policy documents. Empirically, the PRD exhibits a ‘dual-core, core-periphery’ agglomeration pattern centred on Guangzhou and Shenzhen. The cooperation network shows intensifying small-worldness, with density rising from 0.22 to 0.32 and clustering from 0.56 to 0.68 across three observation windows. Six latent communities (modularity Q = 0.41) are anchored by universities and flagship firms. TERGM estimates indicate that non-subsidy policies have a stable positive effect on network optimisation (β = 0.87, 95% CI [0.75, 0.99]), whereas the long-term effect of subsidies turns negative (β = -0.32, 95% CI [-0.59, -0.05]). Enterprise scale, university R&D strength, and geographic proximity promote cooperation, while technological distance inhibits it. Five robustness checks confirm the stability of these results. The findings offer actionable implications for spatial planning, intermediary support, and policy-mix optimisation in the PRD and a reusable analytical template for innovation-chain studies elsewhere.

Short-term passenger-flow forecasting at integrated transport hubs requires accurate forecasts with explicit target semantics. This study evaluates long short-term memory (LSTM) and temporal convolutional network (TCN) fusion models using two distinct 5 min targets at Shenzhen North Station: a 2023–2026 regional population-mobility crowd-density proxy and a 2025 automatic fare collection (AFC) exit-card throughput series. On the proxy task, chronological 70/15/15 splits, a 12-step purge and full-epoch neural training showed that the LSTM-only variant (M2) achieved the lowest neural mean absolute error (MAE) at 5 min, while the fused LSTM–TCN variant (M4) achieved the lowest neural MAE at 15–60 min. Ridge was the strongest conventional proxy reference at 5 min and Random Forest at 15–60 min. On AFC exits, Random Forest was strongest among the six conventional references at all four horizons, while M4 remained better than ridge at 5–60 min. The raw 5 min proxy–AFC Pearson correlation was 0.553; after weekday-by-clock de-seasonalization it was 0.505 (95% day-block confidence interval 0.457–0.538). These values support a shared-signal interpretation; calibration requires a separate mapping model. In the observed-arrival pressure test, recovery at 21:10 reduced cumulative queueing by 67.3%; delays of 10 and 15 min reduced the benefit to 63.4% and 61.7%. The workflow provides 5–60 min lead time for gate and queue preparation, staff rostering and feeder coordination.

In this paper, the concept of reserve capacity has been extended to zone level to measure the land-use development potentiality of each trip generation zone. Bi-level programing models are proposed to determine the signal setting of individual intersections for maximizing possible increase in total travel demand and the corresponding reserve capacity for each zone. The change of the origin–destination pattern with the variation of upper level decision variables is presented through the combined distribution/assignment model under user equilibrium conditions. Both singly constrained and doubly constrained combined models are considered for different trip purposes and data information. Furthermore, we have introduced the continuous network design problem by increasing road capacity and examined its effect on the land-use development potentiality of trip generation zone. A numerical example is presented to illustrate the application of the models and how a genetic algorithm is applied to solve the problem.

Purpose: – The purpose of this paper is to explore the evolving mechanism of urban roadway network. With the consideration of self-organization effect and planning effect during evolution, the authors try to demonstrate the impact of preferential attachment, module scale and module structure on the evolving network model.Design/methodology/approach: – The roadway network is built in the form of abstract network by dual approach. By using the evolving model of modular growth, the authors analyze the effects and mechanism of the evolving process. Then through numerical analysis, the impact of evolving effects on urban roadway network topology structure is discussed from the aspects of preferential attachment, module scale and module structure.Findings: – The module structure property, small-world property and scale-free property of roadway network can be affected with various degrees by the change of preferential attachment and module scale. However, the impact of module structure on network properties is small, which can be ignored. Therefore, in practice, the self-organization effect and planning effect of evolving network can be reached by changing the preferential attachment and module scale, so as to generate the network structure with specific properties.Research limitations/implications: – Some local events, such as road extensions, road demolition and intersection rebuilding, exist during the evolving process under real-world situation. While those cases have not been considered in preferential attachment. Therefore, researchers are encouraged to take these factors into consideration in further research.Practical implications: – The paper has implications for practice in urban transportation planning and roadway constructions, which can help to guide the planning of urban roadway and to adjust or restore partial network when broken down according to the evolving law.Originality/value: – The impact of preferential attachment, module scale and module structure on the evolving network model is measured. And the relationship between different network properties can be used to build some patterns of network. From this point of view, the development of urban roadway network can be predicted and intervened.

Purpose: – By aiming at defending cascade failures effectively, the purpose of this paper is to present a strategy of inserting modular topologies into urban road network through reducing the burdens of critical components with too much traffic flow. Design/methodology/approach: – Each module is considered as a small-world random network, which is inserted into the initial Barabási-Albert scale-free network. Based on the user-equilibrium assignment, the strategy searches for remote nodes with low betweeness and flow in the network, and sets these nodes to be connected with the modular topologies. In this sense, the inserted modules are supposed to attach to the nodes with lower intensity of shorter path, and avoid bringing more impact to the nodes with higher betweeness and traffic flow. By using efficiency as the measurement of cascading failures, the performance of the networks generated through the strategy is tested. Findings: – The results show that the performance of the strategy is sensitive to the average degree of the inserted modular, and the modular size with a better effect on reducing the size of cascading failures or delaying the time of breakdown, while the other factors (e.g. the rewired probability) present few differences among various values. Meanwhile, it is found that the importance-based attachment mechanism has a better effect on preventing the cascading failures, especially delaying the step time of the larger reduction. Practical implications: – The strategy aims at alleviating the burdens in critical components to prevent the cascading failures of the network, and provides practical guidance on the decision of the urban road network evolving process. Originality/value: – An effective strategy for cascade defense in urban road network is proposed in this paper.