We explore neuro-symbolic approaches to generalize actionable knowledge, enabling embodied agents to tackle complex tasks more effectively in open-domain environments. A key challenge for embodied agents is the generalization of knowledge across diverse environments and situations, as limited experiences often confine them to their prior knowledge. To address this issue, we introduce a novel framework, NeSyC, a neuro-symbolic continual learner that emulates the hypothetico-deductive model by continually formulating and validating knowledge from limited experiences through the combined use of Large Language Models (LLMs) and symbolic tools. Specifically, we devise a contrastive generality improvement scheme within NeSyC, which iteratively generates hypotheses using LLMs and conducts contrastive validation via symbolic tools. This scheme reinforces the justification for admissible actions while minimizing the inference of inadmissible ones. Additionally, we incorporate a memory-based monitoring scheme that efficiently detects action errors and triggers the knowledge refinement process across domains. Experiments conducted on diverse embodied task benchmarks—including ALFWorld, VirtualHome, Minecraft, RLBench, and a real-world robotic scenario—demonstrate that NeSyC is highly effective in solving complex embodied tasks across a range of open-domain environments.

Recent advances in large language models (LLMs) have enabled the automatic generation of executable code for task planning and control in embodied agents such as robots, demonstrating the potential of LLM-based embodied intelligence. However, these LLM-based code-as-policies approaches often suffer from limited environmental grounding, particularly in dynamic or partially observable settings, leading to suboptimal task success rates due to incorrect or incomplete code generation. In this work, we propose a neuro-symbolic embodied task planning framework that incorporates explicit symbolic verification and interactive validation processes during code generation. In the validation phase, the framework generates exploratory code that actively interacts with the environment to acquire missing observations while preserving task-relevant states. This integrated process enhances the grounding of generated code, resulting in improved task reliability and success rates in complex environments. We evaluate our framework on RLBench and in real-world settings across dynamic, partially observable scenarios. Experimental results demonstrate that our framework improves task success rates by 46.2\% over Code as Policies baselines and attains over 86.8\% executability of task-relevant actions, thereby enhancing the reliability of task planning in dynamic environments.

In this paper, we target to figure out an adaptive and efficient mechanism that deploys roadside units (RSUs), which are the major components of the vehicle-to-infrastructure (V2I) communication, so that vehicles on the road can have satisfactory connectivity to the vehicular communication infrastructure in a cost-effective way. Only with the conventional fixed RSUs (fRSUs), the coverage requirement of vehicles from the perspective of V2I communications cannot be easily met because of high vehicular mobility compared to the static nature of fRSUs and the high deployment and operational expenditure of fRSUs. Recently, mobile RSUs (mRSUs) mounted on public and commercial vehicles like buses are considered as the substitutes or the supplements of fRSUs. The research on mRSUs is in its early stage and mostly focuses on the deployment of mRSUs from a static viewpoint. In this paper, we consider the environment with densely deployed mRSUs, such as the city environment, thanks to low cost, in which multiple active mRSUs generate lots of control messages to form the mRSU backbone network. To overcome this inefficiency, we propose a mechanism in which each mRSU adaptively and effectively determines its own state, active or inactive, according to the states of its neighboring mRSUs and vehicles. For the NP-hardness proof, we formulate the problem as a 0-1 integer linear programming problem. We evaluate the performance of our mechanism in terms of the ratio of vehicles covered by active mRSUs and the control message overhead compared with the case of nonadaptive mRSU configuration (i.e., the case of all mRSU being active).

Routing protocols for mobile ad hoc networks (MANETs) have assumed that all nodes voluntarily participate in forwarding others' packets. This was a reasonable assumption because all nodes in previous MANETs belonged to a single authority. In the near future, however, a MANET may consist of nodes that belong to multiple organizations since various civilian applications are expected to appear. In this situation, mobile nodes may deliberately avoid packet forwarding to save their own energy, resulting in network partition. To make nodes volunteer in packet forwarding we suggest a payment scheme called protocol-independent fairness algorithm (PIFA). In this type of schemes, nodes need credits to transmit their own packets, and this credit can be earned by forwarding others' packets. PIFA can be utilized irrespective of the type of basic routing protocols, while previous methods are compatible only with source routing like DSR. Simulation results show that PIFA can prevent network partition and network performance degradation due to selfish nodes by inducing them to participate in forwarding packets.

Ad hoc routing protocols such as AODV and DSR do not try to search for new routes until the network topology changes. Hence, with low node mobility, traffic may concentrate on just a few nodes, which results in long end-to-end delay due to congestion at the nodes. Furthermore, since some specific nodes are used for a long time, their battery power may be rapidly exhausted. To resolve this traffic concentration problem, we propose a load balancing scheme called simple load-balancing approach (SLA) that allows each node to drop RREQ or to give up packet forwarding depending on its own traffic load. Meanwhile, mobile nodes may deliberately give up packet forwarding to save their own energy. To make nodes volunteer in packet forwarding we also suggest a payment scheme called protocol-independent fairness algorithm (PIFA).

In most existing studies of wavelength-division multiplexing networks, the problems of routing and wavelength assignment are generally treated separately, since it is NP-complete to produce the optimal solutions for the two problems at the same time. The four presented adaptive routing algorithms, however, consider the availability of wavelengths during the routing process. Our algorithms favor paths with the near-maximum number of available wavelengths between two nodes, resulting in improved load balancing. Simulations show that our algorithms reduce call blocking by nearly half when compared with the least-loaded and the k-fixed routing algorithms in some small networks using the first-fit wavelength assignment policy. In addition, simulation and analysis show that the path length of our algorithms is almost the same as those of the other algorithms.

This paper proposes a new optical path restoration method for wavelength division multiplexing (WDM) networks. In the method, the protection path is established by connecting several sub-lightpaths from the source node to the destination node of the original working lightpath, as opposed to conventional path restoration where a single protection lightpath between the source-destination pair performs restoration. The set of one or more consecutive lightpaths to connect a source-destination pair is called semi-lightpath. The semi-lightpath based restoration provides enhanced flexibility in protection path provisioning and relieves the cost of spare capacity reservation and wavelength conversion at intermediate optical cross-connect nodes. In terms of spare capacity utilization, our method shows substantial reduction of spare capacity overhead compared with dedicated path restoration in all optical networks without wavelength conversion, and shows similar capacity efficiency compared with shared path restoration in opaque networks with full wavelength conversion capability.

This paper proposes a backup network planning method for survivable WDM mesh networks. The proposed method centers around multiple backup cycles where each network link is assigned m backup cycles and each cycle protects 1/m of the working capacity of a target link. Distributed link restoration is performed using preplanned cycles, in which both the backup paths and the spare capacity can be shared. The preconfiguration of the cycles and the spare capacity placement are derived directly from the network topology off-line, which is independent of the primary traffic status or its dynamic changes over time. The proposed method provides efficiency and simplicity to survivable network design and management, and also to runtime recovery operation. Experimental results show that the proposed method needs on average under 60% of spare capacity redundancy for single link failure while preserving the speed of cycle-based restoration.

In this paper, we present an algorithm for efficiently aggregating link state information needed for source routing in PNNI networks. In this algorithm, each border node in a peer group is mapped to a node of a shufflenet. By this mapping, the number of links for which state information is maintained becomes pN (p is an integer, N is the number of border nodes) which is significantly smaller than N/sup 2/ in the full-mesh approach. Another novel aspect of our algorithm is that it can be applied to asymmetric networks, while many previous algorithms such as the spanning tree approach can be applied only to symmetric networks. Experimental results show that our shufflenet algorithm performs as good as the full-mesh approach, with only a much smaller amount of information.

Survivability is one of the important issues in broadband networks, since even a single network element failure may cause a serious amount of data loss. There have been two main approaches for connection restoration: dynamic approach and backup approach. We discuss the trade-offs between the two approaches and propose a new restoration mechanism that centers around the concept of the virtual backup network, which provides backup path sharing as well as spare bandwidth sharing. The proposed mechanism can take advantage of the two existing approaches together with special interests in resource efficiency and routing flexibility.

For many multimedia multicast applications, especially those requiring real-time traffic, it is important that the maximum delay bound requirement between any pair of group members be satisfied. Most studies on multicast routing have been based on a single multicast tree (a shortest path tree or a minimal Steiner tree) approach. By using only a single multicast tree for a multicast group, satisfying the maximum delay bound requirement is nearly impossible. To fulfill these requirement, we adopt the multiple multicast tree concept whose disadvantage is the high tree maintenance cost. Since the tree maintenance cost is proportional to the number of multicast trees for a multicast group, it is necessary to minimize the number of multicast trees. In this paper, we formulate the delay-bounded multicast tree (DBMT) problem whose main objectives are to satisfy an application's maximum delay bound among the group members and to minimize the number of multicast trees needed for a group communication. We categorize the DBMT problem according applications' needs into two subproblems, the shortest path tree-based DBMT (SPT DBMT) and the minimal Steiner tree-based DBMT (MST DBMT) problems. For the DBMT subproblems, we prove the NP-hardness and propose heuristic algorithms. For the performance analysis, simulations were performed.

The paper examines the efficient layout of virtual paths (VPs) in an ATM network. The ATM network consists of ATM switches and their attached network end users, which may be gateways, routers, and hosts. The physical topology, the offered traffic, and call setup matrices of the network end users are assumed to be given. The problem is formulated as a flow-based optimization problem. A heuristic approach is presented which (i) establishes VPs according to physical network-specific and application-specific constraints and a cost function, (ii) provides multipaths between each source destination user pair to minimize the cell blocking probability and to increase network resilience, and (iii) uses a novel VP combining process which is guaranteed to always satisfy the switching constraints. Simulation results are presented for the proposed VP planning policy. Guidelines for the design of robust VP layouts and the efficient establishment of VCs are also presented.>

The authors formulate and propose an algorithm for the LBMT (load-balancing multicast tree) problem whose main objective is to accomplish traffic load balance while minimizing the number of multicast trees for a group. In order to fulfill these objectives, the multiple multicast tree concept, whose only disadvantage is the high tree maintenance cost, is introduced. Since the tree maintenance cost is proportional to the number of multicast trees for a multicast group, it is necessary to minimize the number of multicast trees. The authors' LBMT algorithm is based on the minimal Steiner tree approach by using the information on the available capacities of the links. For the algorithm, several link cost functions which take the available capacities on both directions into account are proposed.