In the field of image encryption, with the continuous development of attack techniques, the requirements for the attack resistance and robustness of encryption algorithms are increasing, and the method of image encryption using a single chaotic system can no longer provide sufficient security and randomness. In this paper, an image encryption method combining S-box and hyperchaotic integrated system is proposed. First, a particle swarm optimization algorithm is designed to search for S-boxes based on an adaptive function that takes into account several security metrics and an S-function inertia weighting strategy. The S-box obtained from the search has good performance index, in which the nonlinearity is 110, the differential uniformity is 6, the algebraic number is 7, and the strict avalanche criterion coefficient is 0.50415. Further, a hyperchaotic integrated system is constructed by combining the Lorenz system and the Henon system with the hyperchaotic Chen system as the core. A three-step diffusion mechanism is adopted in the encryption process, and the S-boxes obtained from the search of the particle swarm optimization algorithm are fully integrated with the chaotic sequences generated by the hyperchaotic integrated system in the first two steps of diffusion to improve the security and randomness of the encryption algorithm. The performance of the proposed encryption algorithm is evaluated and the average values of the calculated metrics are as follows: information entropy value is 7.993, correlation coefficient is 0.008803, NPCR is 99.61%, and UACI is 33.48%. Therefore, the proposed image encryption algorithm combining S-boxes and a hyperchaotic integrated system has high security and robustness, making it applicable to the field of highly demanding image encryption.

This study proposes a multi-dimensional urban transportation network optimization framework (MTNO-RQDC) to address structural failure risks from aging infrastructure and regional connectivity bottlenecks. Through dual-dataset validation using both the Baltimore road network and PeMS07 traffic flow data, we first develop a traffic simulation model integrating Dijkstra’s algorithm with capacity-constrained allocation strategies for guiding reconstruction planning for the collapsed Francis Scott Key Bridge. Next, we create a dynamic adaptive public transit optimization model using an entropy weight-TOPSIS decision framework coupled with an improved simulated annealing algorithm (ISA-TS), achieving coordinated suburban–urban network optimization while maintaining 92.3% solution stability under simulated node failure conditions. The framework introduces three key innovations: (1) a dual-layer regional division model combining K-means geographical partitioning with spectral clustering functional zoning; (2) fault-tolerant network topology optimization demonstrated through 1000-epoch Monte Carlo failure simulations; (3) cross-dataset transferability validation showing 15.7% performance variance between Baltimore and PeMS07 environments. Experimental results demonstrate a 28.7% reduction in road network traffic variance (from 42,760 to 32,100), 22.4% improvement in public transit path redundancy, and 30.4–44.6% decrease in regional traffic load variance with minimal costs. Hyperparameter analysis reveals two optimal operational modes: rapid cooling (rate = 0.90) achieves 85% improvement within 50 epochs for emergency response, while slow cooling (rate = 0.99) yields 12.7% superior solutions for long-term planning. The framework establishes a new multi-objective paradigm balancing structural resilience, functional connectivity, and computational robustness for sustainable smart city transportation systems.