Outcome-based reinforcement learning has made notable advances in training language models (LMs) for reasoning. However, without explicit incentives and controls, this paradigm has limitations and instability in eliciting high-quality reasoning trajectories with diverse actions—particularly for models whose pretraining lacked extensive reasoning-related data. To this end, we introduce MetaAct-RL, a new RL framework that frames LMs’ thinking as sequential decision making over meta-actions. In this framework, the model chooses and executes a high-level action at each step—such as forward reasoning, critique, or refinement—to gradually reach the correct answer. To encourage deeper exploration, richer action diversity, and to improve sampling efficiency in the RL optimization process, MetaAct-RL incorporates appropriate length-based reward and regularization, and a key-state restart mechanism. Extensive experiments across six benchmarks show that MetaAct-RL improves reasoning performance by 7.99 on Llama3.2-1B and 7.17 on Llama3.1-8B relative to vanilla RL method. Moreover, on the challenging AIME-2024, our method outperforms the vanilla RL by 7.5 with Qwen2.5-1.5B.

With the explosive growth of user-generated content (UGC) in multimedia, multimodal sentiment recognition (MSR) tasks face unprecedented challenges. Specifically, existing MSR models exhibit primary limitations: i) Most approaches rely on predefined cross-modal emotion interaction patterns derived from statistical methods, which may not be applicable to all real-world scenarios; ii) many techniques interpret emotions semantically, hindering the ability to capture the sparsity of emotional transmission; and iii) these techniques overlook discrepancies between emotional expressions in images and text on social media. To address these challenges, we developed a novel cross-modal inconsistency adaptive graph learning network, termed RoseNet.RoseNet employs a token-level image-text mapping relationship (ITTM) to effectively capture cross-modal emotional interaction patterns between images and text. Inspired by the Expectation-Maximization (EM) algorithm, and RoseNet features Graph Learning(GL) and Graph Attention prediction(GAP) modules, which are alternately trained to generate more realistic ITTM. Moreover, a carefully designed loss function ensures the sparsity of emotional transmission. Additionally, RoseNet introduces a semantic-affective contrastive learning strategy that enables more grounded semantic and affective representation learning. Evaluation experiments using two benchmarks derived from real-world social media datasets demonstrate that RoseNet achieves significant and consistent performance in MSR tasks. Furthermore, visualizations confirm the model’s ability to capture ITTM with minimal prior knowledge.

The migration of containerized microservices from virtual machines (VMs) to cloud data centers has become the most advanced deployment technique for large software applications in the cloud. This study investigates the scheduling of instance-intensive workflow (IWF) tasks to be executed in containers on a hybrid cloud when computational resources are limited. The process of scheduling these IWF tasks becomes complicated when considering the deployment time of containers, inter-task communication time, and their dependencies simultaneously, particularly when the task can choose multi-mode executions due to the flexible computational resource allocation of the container. We propose a batch scheduling strategy (BSS) for the IWF task scheduling problem. The BSS prioritizes the execution of IWF tasks with high repetition rates with a certain probability and records the virtual machines and modes selected for task execution, which can reduce the data transfer time and the randomness of computation. Based on this, we use an improved hybrid algorithm combined with BSS to solve the multi-mode IWF task scheduling problem. The experimental results demonstrate that employing the BSS can reduce the scheduling time by 6% when the number of workflows increases to 80. Additionally, we tested the effectiveness of all operators in the algorithm, and the results show that each step of the algorithm yields good performance. Compared to similar algorithms in related studies, the overall algorithm can achieve a maximum reduction of approximately 18% in the target value.

Providing potential next point-of-interest (POI) suggestions for users has become a prominent task in location-based social networks, which receives more and more attention from the industry and academia and it remains challenging due to highly dynamic and personalized interactions in user movements. Currently, state-of-the-art works develop various graph- and sequential-based learning methods to model user-POI interactions and transition regularities. However, there are still two significant shortcomings in these works: (1) ignoring personalized spatial and temporal-aspect interactive characteristics capable of exhibiting periodic interests of users and (2) insufficiently leveraging the sequential patterns of interactions for beyond-pairwise high-order collaborative signals among users’ sequences. To jointly address these challenges, we propose a novel multi-view hypergraph learning with spatial-temporal periodic interests for next POI recommendation (MvStHgL). In the local view, we attempt to learn the POI representation of each interaction via jointing periodic characteristics of spatial and temporal aspects. In the global view, we design a hypergraph by regarding interactive sequences as hyperedges to capture high-order collaborative signals across users, for further POI representations. More specifically, the output of POI representations in the local view is used for the initialized embedding, and the aggregation and propagation in the hypergraph are performed by a novel Node-to-Hypergraph-to-Node scheme. Furthermore, the captured POI embeddings are applied to achieve sequential dependency modeling for next POI prediction. Extensive experiments on three real-world datasets demonstrate that our proposed model outperforms the state-of-the-art models.

Motivated by rising concerns regarding global warming and traffic congestion effects, we study the time-dependent green vehicle routing problem with time windows (TDGVRPTW), aiming to minimize carbon emissions. The TDGVRPTW is a variant of the time-dependent vehicle routing problem (TDVRP) in which, in addition to the time window constraints, the minimization of carbon emissions requires determination of the optimal departure times for vehicles, from both the depot and customer location(s). Accordingly, the first exact method based on a branch-cut-and-price (BCP) algorithm is proposed for solving the TDGVRPTW. We introduce the notation of a time-dependent (TD) arc and describe how to identify the nondominated TD arcs in terms of arc departure times. In this way, we reduce infinitely many TD arcs to a finite set of nondominated TD arcs. We design a state-of-the-art BCP algorithm for the TDGVRPTW with labeling and limited memory subset row cuts, together with effective dominance rules for eliminating dominated TD arcs. The exact method is tested on a set of test instances derived from benchmark instances proposed in the literature. The results show the effectiveness of the proposed exact method in solving TDGVRPTW instances involving up to 100 customers. Summary of Contribution: Due to the environmental situation, green vehicle routing problems (GVRPs) aim to consider greenhouse gas emissions reduction, while routing the vehicles, and play a key role in transportation and logistics. Vehicle greenhouse gas emissions strongly depend on the vehicle speeds and traffic conditions which in real life vary continuously over time. To tackle these challenges, we address the time-dependent green vehicle routing problem with time windows (TDGVRPTW) aimed at reducing total carbon emissions under time-dependent travel times and time window constraints. We design an effective exact method for the TDGVRPTW based on a state-of-the-art branch-cut-and-price algorithm. The paper is both of methodological value for researchers and of interest for practitioners. For researchers, the presented algorithm is amenable for various routing constraints and provides a ground for further studies and research. For practitioners, the paper suggests insights on how the carbon emissions change based on different vehicle speed profiles. History: Accepted by Andrea Lodi, Area Editor for Design & Analysis of Algorithms—Discrete. Funding: This research is supported by the National Natural Science Foundation of China [Grants 71831003, 71831006, 72171043, and 71901180] and the Fundamental Research Funds for the Central Universities [Grants N170405005 and N180704015]. Supplemental Material: The electronic companion is available at https://doi.org/10.1287/ijoc.2022.1195 .

Abstract Bullet screen comments (BSCs) are user‐generated short comments that appear as real‐time overlays on many video platforms, expressing the audience opinions and emotions about different aspects of the ongoing video. Unlike traditional long comments after a show, BSCs are often incomplete, ambiguous in context, and correlated over time. Current studies in sentiment analysis of BSCs rarely address these challenges, motivating us to develop an aspect‐level sentiment analysis framework. Our framework, BSCNET, is a pre‐trained language encoder‐based deep neural classifier designed to enhance semantic understanding. A novel neighbor context construction method is proposed to uncover latent contextual correlation among BSCs over time, and we also incorporate semi‐supervised learning to reduce labeling costs. The framework increases F1 (Macro) and accuracy by up to 10% and 10.2%, respectively. Additionally, we have developed two novel downstream tasks. The first is noisy BSCs identification, which reached F1 (Macro) and accuracy of 90.1% and 98.3%, respectively, through fine‐tuning the BSCNET. The second is the prediction of future episode popularity, where the MAPE is reduced by 11%–19.0% when incorporating sentiment features. Overall, this study provides a methodology reference for aspect‐level sentiment analysis of BSCs and highlights its potential for viewing experience or forthcoming content optimization.

Solutions to the weighted vehicle routing problem (WVRP) find numerous applications, such as home-to-work bus service, toll-by-weight cargo transportation, perishable-food delivery, and hazardous waste collection. This paper presents mathematical WVRP models corresponding to both collection and delivery cases. This study demonstrates WVRPs to be NP-hard, and proposes an efficient heuristic method (RI-ILS) to solve WVRPs. RI-ILS is based on principles of regret-insertion and iterative local-search. Several computational experiments were performed in this study to demonstrate the utility of the RI-ILS approach. As observed, when solving traditional vehicle routing problems (VRPs), RI-ILS yields good results under all testing conditions. In particular, it provides new best-known solutions under 12 testing conditions. When solving WVRPs, the RI-ILS method outperforms two recently published state-of-the-art algorithms in terms of solution quality, computational time, and stability. Case studies have also been performed in this research based on real-world data obtained from two Chinese companies that provide home-to-work bus service to their employees.

In real-world cargo transportation, there are charges associated with both the traveling distance and the loading quantity. Cargo trucks must comply with a mandatory lower carbon emissions policy: the emissions of large-volume cargo truck/containers depend greatly on the cargo loading and the traveling distance. To address this issue, instead of assuming a constant vehicle loading from one customer to another, a variable vehicle loading should be used in optimizing the vehicle routine, which is known as a weighted vehicle routing problem (WVRP) model. The WVRP is an NP-hard problem; thus, the purpose of this paper is to develop a BEAM-MMAS algorithm that combines a MAX-MIN ant system with beam search to show that the WVRP is more effective than the VRP and to determine the types of VRP instances for which the WVRP has more cost-savings than the VRP. To this end, computational experiments are carried out on benchmark problems of the capacitated VRP for seven types of distributions, and the effectiveness of the BEAM-MMAS algorithm is compared with that of general ACO and MMAS algorithms for large-size benchmarking instances. The benchmarking tests show that lower operation costs are produced using the WVRP than using the optimal or best known paths of the CVRP and that the WVRP can increase cost savings for the instances with a dispersed customer distribution and a large weight.

Cell formation and parts scheduling are two important correlated processes in a cellular manufacturing system; however, the decisions involved in these processes are typically made individually. Determining how to integrate these decisions effectively to pursue a productive and lower cost system has become an important issue. This paper focuses on providing an effective solution to integrate the decisions of cell formation and parts scheduling, while considering intercell moves by using a Lagrangian relaxation decomposition method. A mixed integer nonlinear programming mathematical model (CFPSP) is proposed to determine which part families and machine groups are assigned to cells and in which sequence the parts are processed in the machines to minimize the total tardiness penalty cost. To effectively solve the model, a Lagrangian relaxation decomposition method with a heuristic (LRDH) is developed. Using the LRDH, the CFPSP model is solved by decomposing the model into two subproblems, i.e., the cell formation subproblem (CFPSP-FD) and the parts scheduling subproblem (CFPSP-SD). After linearizing the CFPSP-FD model, the subproblem CFPSP-FD is solved by the MIP optimizer CPLEX. A scatter search approach is developed to solve the subproblem CFPSP-SD. Combined with the Lagrange multipliers, the CFPSP-SD model takes into consideration the assignment of part families and the associated machine groups to each cell, when it sequences the processing of the parts on each machine in cells. An illustration of the application of the CFPSP model in an electronic appliance cellular manufacturing enterprise in China is presented.

Quality function deployment (QFD) is a well-known customer driven system and tool for planning and managing new product development. House of quality (HOQ) is used to translate customer requirements (CRs) into each stage of a product development in QFD. In the conventional QFD model, constant coefficients are employed to present the relationships between “What matrix” and “How matrix” and among “How matrixes". In order to break the hindrance of the representation of simple linear function in the structure of HOQ, an integrative approach considering Kano model into QFD planning are employed, and processes are performed in uncertain and vague environments. Therefore, a fuzzy multi-objective model (FMOQFD-Kano) is proposed to reconcile the tradeoff between customer satisfaction and cost. Not only customer satisfaction is taken into account, but also cost is considered in the proposed model. Moreover, a fuzzy solution approach is applied to solve the model for preferred solutions under different alpha-levels, which can give decision-maker more choices to cater for uncertain environment. An example is presented to illustrate the applicability of the proposed model.

Given the uncertain market demands and capacities in production environment, this paper discusses some practical approaches to modeling multiproduct aggregate production planning problems with fuzzy demands, fuzzy capacities, and financial constraints. By formulating the fuzzy demand, fuzzy equation, and fuzzy capacities, a fuzzy production-inventory balance equation for single period and a dynamic balance equation are formulated as fuzzy/soft equations and they represent the possibility levels of meeting the market demands. Using this formulation and interpretation, a fuzzy multiproduct aggregate production planning model is developed, and its solutions using parametric programming, best balance and interactive techniques are introduced to cater to different scenarios under various decision making preferences. Using the proposed models and techniques, first, the decision maker can select a preferred production plan with a common satisfaction level or different combinations of preferred possibility level and satisfaction levels, according to the market demands and available production capacities, and second, the obtained structure of the optimal solution can help decision maker in aggregate production planning. The decision maker can also make a preferred and reasonable production plan corresponding to one's most concerned criteria. Hence, decision makers not only can come up with a reasonable aggregate production plan with minimum efforts, but also have more choices of making a preferred aggregate plan based on his most concerned criteria. These models can effectively enhance the capability of an aggregate plan to give feasible family disaggregation plans under different scenarios with fuzzy demands and capacities. Simulation and the results of analysis on the proposed techniques are also given in detail in this paper.