This paper addresses the challenges of topology control, multi-hop routing, and link mode selection in optical-acoustic hybrid underwater wireless sensor networks (OA-UWSNs) by proposing a joint optimization framework that integrates hierarchical reinforcement learning (HRL) with an improved non-dominated sorting genetic algorithm (NSGA-III+). The OA-UWSN is formulated as a multi-objective graph optimization problem subject to connectivity, mode allocation, and reliability constraints. At the topology and link layers, the HRL-based approach employs a deep deterministic policy gradient (DDPG) algorithm at the upper level to assign connectivity weights to network links and incorporates a minimal connectivity restoration mechanism to construct feasible subgraphs. At the lower level, Q-learning is adopted to enable adaptive selection between acoustic and optical transmission modes for each link. For the routing layer, NSGA-III+ is developed to achieve multi-objective route optimization, leveraging segmented path encoding, a dynamically weighted fitness function, and a multidimensional crowding adjustment strategy to enhance both the convergence rate and the diversity of the solution set. Simulation results demonstrate that the proposed HRL+NSGA-III+ framework consistently outperforms existing baseline methods in terms of convergence speed, energy efficiency, latency reduction, and link quality, and exhibits robust performance across various network scales.

With the increasing number of terminal devices in the Ocean of Things (OoT), it is necessary to apply the OoT mobile edge computing (MEC) paradigm to low-Earth orbit (LEO) satellites. The aim is to support the operation of compute-intensive OoT services with LEO satellite assistance. To address the proliferation of computing services in OoT, this article proposes a satellite-assisted task offloading and resource allocation (STORA) approach for OoT edge computing, which includes a generalized framework for three-layer MEC systems in space, on the surface, and underwater. First, the MEC system energy minimization problem is described as mixed integer-nonlinear programming (MINLP) and divided into two subproblems: 1) task offloading and 2) resource allocation. Second, the task offloading subproblem is modeled as a Markov decision process (MDP). The proposed adaptive deep deterministic policy gradient (A-DDPG) algorithm jointly optimizes the offloading policy and offloading volume. In A-DDPG, a soft network update method with an adaptive updating coefficient ensures stable network updates while achieving fast convergence. Finally, the resource allocation is decomposed into a joint optimization problem involving buoy and satellite computational resources, which is shown to be convex. The Lagrange multiplier method is used to optimize the buoy-satellite resource allocation problem while also balancing edge computational load across servers. The experimental results show that STORA can reduce network energy consumption by 17.8%, increase network lifetime by 24.4%, and lower network latency by 11.5%.

To build the new generation of ubiquitous communication and service integration networks with “network omnipresence and computing ubiquitous,” it is urgent to improve the in-orbit computing ability of low earth orbit (LEO) constellation networks and develop intelligent technology for satellite–ground collaborative edge computing. Communication tasks between ground nodes and satellites are increasing, but the satellite-to-ground spectrum resources are limited. The reasonable application of channels determines the performance of the network, which in turn affects the users’ experience. An outstanding issue is how to allocate channels rationally and control the power of data transmission to reduce co-channel interference and minimize system overhead effectively. This article studies multiuser computation offloading for low earth orbit (LEO) constellation networks under dynamic environment, wherein the system overhead is minimized by joint offloading strategy and power optimization. First, we propose a generic network architecture for computation offloading of LEO constellation networks under the dynamic environment. Then, from a game-theoretic perspective, we model the overhead minimization problem as a potential game and prove that the Nash equilibrium (NE) minimizes the system overhead. After that, to reach the NE, we design the Synchronous log-linear learning-based power control algorithm and joint offloading strategy and power optimization algorithm based on SLA (JOPAS), and prove the convergence of the algorithms. Finally, the effectiveness of the proposed algorithm is verified through extensive simulations and comparisons with benchmark algorithms, and the proposed algorithm achieves near-optimal performance.

This paper introduces a protection mechanism for complete recovery from multi-link failures in Low Earth Orbit (LEO) satellite networks. Inter-satellite links (ISLs) may break down frequently due to the interruption of wireless channels, the high-speed movement of satellites and the long distance for communication. As a proactive scheme for recovery, IP Fast Reroute (IPFRR) has been used to achieve fast rerouting via pre-calculating backup paths. Although IPFRR has effectively addressed the problem of single-link failures, the recovery from multi-link failures still needs further studies, especially on aspects such as recovery rate, computation complexity and storage cost. For better survivability of LEO satellite networks, we develop an IPFRR mechanism for recovering from multi-link failures. This new mechanism, called Grid Bypass Routing (GBR), assigns two addresses to every network device – a normal address and a protection address. When failures occur, routers will choose the protection address to deliver packets via alternative paths. We develop a technique to compute backup paths and prove that GBR can guarantee complete recovery. We evaluate GBR in the simulation, and the result shows that the advantage of its computational cost, forwarding table size and recovery rate makes GBR a more efficient and easily manageable IPFRR solution to link failures in LEO satellite networks.

Internet of Space Things (IoST), which extends the concept of Internet of Things (IoT) to space, has emerged as a new paradigm for offering monitoring/reconnaissance, in-space backhaul, and cyber-physical integration services. As Low Earth Orbit (LEO) satellites are increasingly deployed for global Internet services, Mobile Edge Computing (MEC) is being introduced into satellite networks to provision computing services by placing edge servers on satellites. Nevertheless, it is a nontrivial and unexplored task to efficiently choose edge server deployment locations from a large number of satellites. In this paper, we address this issue in detail towards average response delay minimization considering propagation delay, forwarding delay, and service delay. In particular, we formulate the dynamic server placement problem as well as the static server placement problem, and devise a genetic algorithm-based heuristic approach to solve them. Simulation results compare the two placement strategies with two benchmarks and demonstrate the performance of our genetic algorithm-based approach. Furthermore, a comparison between the dynamic placement and the static placement is investigated.

The increasingly extensive availability of location- acquisition technologies (such as GPS and GSM networks) and mobile computing techniques have generated a lot of spatial-temporal trajectory data which represents the mobility of diversification of moving objects such as people, vehicles, and animals. This brings new opportunities to understand the movement behaviors of moving objects in mobile social networks, which can become valuable for location-based services (LBS). However, the explosive growth of users' trajectory data has brought a lot of troubles to online real-time processing. In this paper, we focus on this direction and develop a complete data-driven framework involving real-time user mobility pattern modeling and a novel user similarity measurement based on both spatial and temporal information. Through experimental evaluation, it is verified that the proposed mobility pattern modeling method and similarity measurement can deliver excellent performance.