The multiple-choice knapsack problem (MCKP) is a classic NP-hard combinatorial optimization problem. Motivated by several significant real-world applications, this work investigates a novel variant of MCKP called the chance-constrained MCKP (CCMCKP), where item weights are random variables. In particular, we focus on the practical scenario of CCMCKP, in which the probability distributions of random weights are unknown and only sample data is available. We first present the problem formulation of CCMCKP and then establish the two benchmark sets. The first set contains synthetic instances, while the second set is designed to simulate a real-world application scenario of a telecommunication company. To solve CCMCKP, we propose a data-driven adaptive local search (DDALS) algorithm. Compared to existing stochastic optimization and distributionally robust optimization methods, the main novelty of DDALS lies in its data-driven solution evaluation approach, which does not make any assumptions about the underlying distributions and is highly effective even when faced with a high intensity of the chance constraint and a limited amount of sample data. Experimental results demonstrate the superiority of DDALS over the baselines on both the benchmarks. Finally, DDALS can serve as the baseline for future research, and the benchmark sets are open-sourced to further promote research on this challenging problem.

Federated Learning (FL) is a recent development in distributed machine learning that collaboratively trains models without training data leaving client devices, preserving data privacy. In real-world FL, the training set is distributed over clients in a highly non-Independent and Identically Distributed (non-IID) fashion, harming model convergence speed and final performance. To address this challenge, we propose a novel, generalised approach for incorporating adaptive optimisation into FL with the Federated Global Biased Optimiser (FedGBO) algorithm. FedGBO accelerates FL by employing a set of global biased optimiser values during training, reducing ‘client-drift’ from non-IID data whilst benefiting from adaptive optimisation. We show that in FedGBO, updates to the global model can be reformulated as centralised training using biased gradients and optimiser updates, and apply this framework to prove FedGBO's convergence on nonconvex objectives when using the momentum-SGD (SGDm) optimiser. We also conduct extensive experiments using 4 FL benchmark datasets (CIFAR100, Sent140, FEMNIST, Shakespeare) and 3 popular optimisers (SGDm, RMSProp, Adam) to compare FedGBO against six state-of-the-art FL algorithms. The results demonstrate that FedGBO displays superior or competitive performance across the datasets whilst having low data-upload and computational costs, and provide practical insights into the trade-offs associated with different adaptive-FL algorithms and optimisers.

Multi-access Edge Computing (MEC) is an emerging computing paradigm that extends cloud computing to the network edge to support resource-intensive applications on mobile devices. As a crucial problem in MEC, service migration needs to decide how to migrate user services for maintaining the Quality-of-Service when users roam between MEC servers with limited coverage and capacity. However, finding an optimal migration policy is intractable due to the dynamic MEC environment and user mobility. Many existing studies make centralized migration decisions based on complete system-level information, which is time-consuming and also lacks desirable scalability. To address these challenges, we propose a novel learning-driven method, which is user-centric and can make effective online migration decisions by utilizing incomplete system-level information. Specifically, the service migration problem is modeled as a Partially Observable Markov Decision Process (POMDP). To solve the POMDP, we design a new encoder network that combines a Long Short-Term Memory (LSTM) and an embedding matrix for effective extraction of hidden information, and further propose a tailored off-policy actor-critic algorithm for efficient training. The extensive experimental results based on real-world mobility traces demonstrate that this new method consistently outperforms both the heuristic and state-of-the-art learning-driven algorithms and can achieve near-optimal results on various MEC scenarios.

Federated learning (FL) is a privacy-preserving distributed machine learning paradigm that enables collaborative training among geographically distributed and heterogeneous devices without gathering their data. Extending FL beyond the supervised learning models, federated reinforcement learning (FRL) was proposed to handle sequential decision-making problems in edge computing systems. However, the existing FRL algorithms directly combine model-free RL with FL, thus often leading to high sample complexity and lacking theoretical guarantees. To address the challenges, we propose a novel FRL algorithm that effectively incorporates model-based RL and ensemble knowledge distillation into FL for the first time. Specifically, we utilise FL and knowledge distillation to create an ensemble of dynamics models for clients, and then train the policy by solely using the ensemble model without interacting with the environment. Furthermore, we theoretically prove that the monotonic improvement of the proposed algorithm is guaranteed. The extensive experimental results demonstrate that our algorithm obtains much higher sample efficiency compared to classic model-free FRL algorithms in the challenging continuous control benchmark environments under edge computing settings. The results also highlight the significant impact of heterogeneous client data and local model update steps on the performance of FRL, validating the insights obtained from our theoretical analysis.

Edge computing is an emerging promising computing paradigm that brings computation and storage resources to the network edge, hence significantly reducing the service latency and network traffic. In edge computing, many applications are composed of dependent tasks where the outputs of some are the inputs of others. How to offload these tasks to the network edge is a vital and challenging problem which aims to determine the placement of each running task in order to maximize the Quality-of-Service (QoS). Most of the existing studies either design heuristic algorithms that lack strong adaptivity or learning-based methods but without considering the intrinsic task dependency. Different from the existing work, we propose an intelligent task offloading scheme leveraging off-policy reinforcement learning empowered by a Sequence-to-Sequence (S2S) neural network, where the dependent tasks are represented by a Directed Acyclic Graph (DAG). To improve the training efficiency, we combine a specific off-policy policy gradient algorithm with a clipped surrogate objective. We then conduct extensive simulation experiments using heterogeneous applications modelled by synthetic DAGs. The results demonstrate that: 1) our method converges fast and steadily in training; 2) it outperforms the existing methods and approximates the optimal solution in latency and energy consumption under various scenarios.

With the massive deployment of 5G cellular infrastructures, traffic prediction has become an indispensable part of the cellular resource management system in order to provide reliable and fast communication services that can meet the increasing quality-of-service requirements of smart city. A promising approach for handling this problem is to introduce intelligent methods to implement a highly effective and efficient cellular traffic prediction model. Meanwhile, integrating the multiaccess edge computing framework in 5G cellular networks facilitates the application of intelligent traffic prediction models by enabling their implementation at the network edge. However, the data shortage and privacy issues may still be obstacles for training a robust and accurate prediction model at the edge. To address these issues, we propose a data-augmentation-based cellular traffic prediction model (ctGAN-S2S), where an effective data augmentation submodel based on generative adversarial networks is proposed to improve the prediction performance while protecting data privacy, and a long-short-term-memory-based sequence-to-sequence submodel is used to achieve the flexible multistep cellular traffic prediction. The experimental results on a real-world city-scale cellular traffic dataset reveal that our ctGAN-S2S model achieves up to 48.49% improvement of the prediction accuracy compared to four typical reference models.

Multi-access edge computing (MEC) aims to extend cloud service to the network edge to reduce network traffic and service latency. A fundamental problem in MEC is how to efficiently offload heterogeneous tasks of mobile applications from user equipment (UE) to MEC hosts. Recently, many deep reinforcement learning (DRL)-based methods have been proposed to learn offloading policies through interacting with the MEC environment that consists of UE, wireless channels, and MEC hosts. However, these methods have weak adaptability to new environments because they have low sample efficiency and need full retraining to learn updated policies for new environments. To overcome this weakness, we propose a task offloading method based on meta reinforcement learning, which can adapt fast to new environments with a small number of gradient updates and samples. We model mobile applications as Directed Acyclic Graphs (DAGs) and the offloading policy by a custom sequence-to-sequence (seq2seq) neural network. To efficiently train the seq2seq network, we propose a method that synergizes the first order approximation and clipped surrogate objective. The experimental results demonstrate that this new offloading method can reduce the latency by up to 25 percent compared to three baselines while being able to adapt fast to new environments.

Vehicular edge computing (VEC) is a new computing paradigm that has great potential to enhance the capability of vehicle terminals (VTs) to support resource-hungry in-vehicle applications with low latency and high energy efficiency. In this article, we investigate an important computation offloading scheduling problem in a typical VEC scenario, where a VT traveling along an expressway intends to schedule its tasks waiting in the queue to minimize the long-term cost in terms of a tradeoff between task latency and energy consumption. Due to diverse task characteristics, dynamic wireless environment, and frequent handover events caused by vehicle movements, an optimal solution should take into account both where to schedule (i.e., local computation or offloading) and when to schedule (i.e., the order and time for execution) each task. To solve such a complicated stochastic optimization problem, we model it by a carefully designed Markov decision process (MDP) and resort to deep reinforcement learning (DRL) to deal with the enormous state space. Our DRL implementation is designed based on the state-of-the-art proximal policy optimization (PPO) algorithm. A parameter-shared network architecture combined with a convolutional neural network (CNN) is utilized to approximate both policy and value function, which can effectively extract representative features. A series of adjustments to the state and reward representations are taken to further improve the training efficiency. Extensive simulation experiments and comprehensive comparisons with six known baseline algorithms and their heuristic combinations clearly demonstrate the advantages of the proposed DRL-based offloading scheduling method.

Inspired by mobile edge computing (MEC), vehicular edge computing (VEC) enables vehicle terminals to support resource-hungry on-vehicle applications with significantly lower latency and less energy consumption. In this paper, we investigate the computation offloading problem in a typical VEC scenario, where a vehicle offloads its computation tasks to the VEC servers deployed in the road side unit (RSU) to minimize its long-term user cost. The mobility of the vehicle coupled with the high dynamics of the environment makes the problem particularly difficult. To tackle this challenge, a deep reinforcement learning (DRL) based offloading method is proposed, which approximates the offloading policy (OP) by a deep neural network (DNN) and trains the DNN with the proximal policy optimization (PPO) algorithm without a priori knowledge of the environment dynamics. Extensive simulation experiments and comprehensive comparison with six baseline algorithms demonstrate that it can achieve the lowest user cost in most cases.