This paper presents the Smart Local Search (SLS) algorithm, a hybrid metaheuristic framework for the NP-hard examination timetabling problem. SLS distinctively integrates graph-based construction heuristics with a guided local search framework, enhanced by an aspiration criterion to recover promising solutions overlooked by penalty terms and a strategic random move mechanism to escape local optima. Extensive experiments on all 11 of Carter’s widely used benchmark datasets show that SLS achieves competitive results, with an average penalty of 28.03 and a standard deviation of 42.30 across the benchmarks. Statistical significance testing (Wilcoxon signed-rank test, α = 0.05) confirms that SLS’s performance is significantly better than most baseline methods (p-value−10). The algorithm demonstrates particular strength in stability and consistency, making it a robust and efficient solution for complex educational scheduling environments.

The emergence of Cooperative Intelligent Transportation Systems (C-ITS) has revolutionized urban mobility by enabling seamless collaboration among vehicles, infrastructure, and individuals to improve traffic management, safety, and efficiency. Deep Reinforcement Learning (DRL) has become a key technology in this ecosystem, empowering autonomous agents to make real-time decisions that optimize traffic flow, reduce congestion, and enhance road safety. Although many surveys on Intelligent Transportation Systems (ITS) either overlook cooperative aspects or primarily emphasize security, this paper bridges the gap by examining the diverse applications of DRL in C-ITS. It examines critical areas such as traffic signal control, AV coordination, route planning, and human-vehicle interaction. The study also traces the evolution of DRL algorithms, their adaptation to transportation challenges, and their integration with cutting-edge projects and standards. Additionally, the paper provides a comprehensive analysis of current research trends, identifying achievements, unresolved challenges, and future directions in the field. By synthesizing existing literature and highlighting the synergy between DRL and C-ITS, this survey serves as a valuable resource for researchers, policymakers, and industry professionals striving to develop intelligent, cooperative, and sustainable transportation systems. The insights offered aim to guide advancements in this rapidly growing domain, fostering innovation and practical implementation.

The emergence of Autonomous Vehicles has ushered in a new era of transportation efficacy and safety. Platooning, which involves vehicles traveling closely in sync, offers potential for mitigating traffic congestion, decreasing fuel usage, and improving road safety. However, realizing platooning’s full potential requires robust control strategies adaptable to various conditions. This study explores integrating Federated Deep Reinforcement Learning (FDRL) into AV platooning systems to enhance control and efficiency. It focuses on leveraging FDRL to optimize platoon behavior while considering AV’s distributed nature. Specifically, the proposed approach involves training Deep Reinforcement Learning (DRL) model locally on individual vehicles (Agents) within a platoon, allowing them to adapt and learn from local data and experiences. The trained model is then transferred to other platoons via 5 G infrastructure, improving overall performance. Simulation studies demonstrate the superiority of FDRL over other methods, suggesting its potential to advance AV platooning systems in future transportation landscapes.

Abstract Sustainable agriculture is becoming increasingly important in the face of growing environmental challenges. One key aspect of sustainable agriculture is managing soil resources effectively. In this context, digital soil mapping (DSM) has emerged as a powerful tool to understand soil variability better and inform land management decisions. This paper proposes a comprehensive data warehouse for DSM that supports climate change analysis. Our architecture integrates frequent itemset mining (FMI) and association rules mining (ARM) to extract insights from large‐scale soil data. We review related studies in soil data warehousing and ARM, identify gaps, and propose a data warehouse architecture leveraging the galaxy multidimensional model for DSM based on the SCORPAN model, which incorporates all relevant soil forming factors. We employ and compare A‐priori, FP‐growth, and ECLAT algorithms to efficiently mine frequent itemsets and generate association rules. Our intensive experiments evaluation demonstrates that FP‐growth outperforms the other algorithms in accuracy, scalability, and speed and requires less memory. Additionally, we utilized correlation metrics for ARM, such as lift, cosine, kulc, and Imbalance ratio, to obtain the most significant and relevant association rules. These rules provide valuable insights into the complex relationships between soil properties and environmental factors, which can inform land management decisions and improve sustainable agriculture practices. This work contributes to the growing body of research on DSM and data‐driven approaches to sustainable agriculture.

Abstract In this article, the paradigm of machine culture as an extension to machine intelligence is introduced. This new concept is modelled based on animal intelligence and culture. The example of orca intelligence and culture is considered as orcas possess in addition to skills allowing them to reach preys, the ability to transmit their culture from generation to generation. The orca intelligence is studied and then simulated to design an algorithm called Orca Algorithm (OA). OA consists in modelling the orca lifestyle and in particular the orcas social organization, echolocation behaviour and hunting techniques. In order to integrate the cultural dimension, OA was hybridized with the Cultural Algorithm (CA) to get an algorithm called Orca Cultural Algorithm (OCA). OCA was tested on 22 benchmark problems of the literature to evaluate its performance. Extensive experiments were first performed to set the algorithm parameters prior to measure its effectiveness and efficiency. In a second stage, OCA was adapted to discrete problems and applied to the maze game with four level of complexity. Additional experiments were held to compare the designed algorithm with recent state‐of‐the‐art evolutionary algorithms. The overall obtained results are very promising.

Abstract In this paper, we are proposing two novel swarm intelligence approaches for solving discrete problems. Two different continuous evolutionary algorithms inspired by the behaviour of elephants, namely elephant herding optimization and elephant swarm water search algorithm, are studied and analysed in order to propose discrete versions of the latter called discrete elephant herding optimization and discrete elephant swarm water search algorithm. As an illustration of how our proposals can work on discrete problems, a case study on association rule mining is carried out where the proposed discrete algorithms are modelled and applied on the problem in order to extract interesting association rules from large‐scale databases. Extensive experiments on eight different real‐life and relevant datasets with various sizes showed that both our proposals yield good results that compete and sometimes outperform state of the art algorithms.

Summary Detecting hiding communities is considered as a main topic in complex networks. In this article, we propose a multi‐swarm elephant herding optimization (EHO) algorithm to uncover community structures in complex environments. It adapts EHO algorithm to community detection problem. EHO algorithm relies on two procedures which are updating clan procedure and separating procedure. The main idea of our multi‐swarm approach is that the population is composed of a set of interacting clans. In each clan, a local search function is defined to determine best local individual called matriarch. Through updating clan procedure, the remaining individuals in the clan update their positions based on the matriarch position. In addition, to ensure significant individuals in the clan, a multi‐swarm cooperative algorithm is designed to implement separating procedure; clans interchange individuals to balance the exploration and exploitation abilities. A series of experiments are carried out on artificial and real networks. The results obtained by the proposed approach are better than the results obtained by some other approaches.

In the field of data science, we consider usually data independently from a problem to be solved. The originality of this paper consists in handling huge instances of combinatorial problems with datamining technologies in order to reduce the complexity of their treatment. Such task can be performed on Web combinatorial optimization such as internet data packet routing and web clustering. We focus in particular on the satisfiability of Boolean formulae but the proposed idea could be adopted for any other complex problem. The aim is to explore the satisfiability instance using datamining techniques in order to reduce its size, prior to solve it. An estimated solution for the obtained instance is then computed using a hybrid algorithm based on DPLL technique and a genetic algorithm. It is then compared to the solution of the initial instance in order to validate the method effectiveness. We performed experiments on the wellknown BMC datasets and show the benefits of using datamining techniques as a pretreatment, prior to solving the problem.

In this paper, a hybrid system combining ant based approaches and tabu search has been designed for the generation of materialized views in a relational data warehouse environment with the purpose of improving the queries performance. Two ACO algorithms were adapted for the views generation problem to take up the scalability challenge and information retrieval technologies are used in the search process. In addition, our approach manages dynamically the storage to include the best views determined by the bio-inspired approach. Experiments have been conducted to validate the designed algorithms and interesting performance is observed when comparing it with those of the previous related works.

When dealing with large scale applications, data sets are huge and very often not obvious to tackle with traditional approaches. In web information retrieval, the greater the number of documents to be searched, the more powerful approach required. In this work, we develop document search processes based on particle swarm optimization and show that they improve the performance of information retrieval in the web context. Two novel PSO algorithms namely PSO1-IR and PSO2-IR are designed for this purpose. Extensive experiments were performed on CACM and RCV1 collections. The achieved results exhibit the superiority of PSO2-IR on all the others in terms of scalability while yielding comparable quality.

Unlike existing studies dealing with the selection of Bitmap Join Indexes for star join queries optimization, this paper presents three original features. The first one consists in addressing the problem with ant based approach that is more robust than the simple heuristic algorithms, which are usually used in the related works. The second interesting novelty resides in the metric used to prune the search space. The fitness function designed in the ant approach is brought from information retrieval technologies and is more refined than the frequency measure usually used. Finally, the third efficient aspect is in the data structure used to manage dynamically the storage in order to select the best promising indexes.

This paper deals with large scale information retrieval aiming at contributing to web searching. The collections of documents considered are huge and not obvious to tackle with classical approaches. The greater the number of documents belonging to the collection, the more powerful approach required. A Bees Swarm Optimization algorithm called BSO-IR is designed to explore the prohibitive number of documents to find the information needed by the user. Extensive experiments were performed on CACM and RCV1 collections and more large corpuses in order to show the benefit gained from using such approach instead of the classic one. Performances in terms of solutions quality and runtime are compared between BSO and exact algorithms. Numerical results exhibit the superiority of BSO-IR on previous works in terms of scalability while yielding comparable quality.

This paper proposes a new approach using a multi agents system for personalizing the information source selection. Most prior research for information source selection focused on selecting the sources that has the most relevant content according to the query but ignored the user’s specific needs. Our approach extends the state of the art in distributed information retrieval. First, it develops models for representing both user and information source using feature based profiles. Second, it develops an agent called user-agent for managing the user profile. Third, it develops an agent called source-agent for each information source in order to manage its information source (source profile) in parallel. Fourth, it develops an agent called agent-broker for cooperating between user-agent and each source-agent in order to select the best source to the user’s query. The approach has been experimented with several known information sources. The experimental results obtained show that the approach: (1) Improve the relevance of the result. (2) Reduce the response times. (3) Improve the system extensibility.

This paper presents a new parallel evolutionary approach where the concept of generation has been removed and replaced by the cycle one. Indeed, the classical genetic algorithms (GAs) deals with operations on the whole population through all generations. These operations are performed during the evolution towards the best individual or solution of the considered combinatorial problem. In our approach, each individual participates to the evolutionary process uniquely during some iterations. There is no generation where all individuals are created at the same time and disappear at the same time at the end of the evolutionary process genation. In our approach, each individual owns one lifespan represented by a number of cycles which are affected to it randomly at its birth and at the end of which it disappears from the population. Consequently, only certain individuals of the population are evaluated within each iteration of the algorithm and not all the population. This causes the substantial reduction of the total running time of the algorithm since the evaluations of all individuals of each generation necessitates more than 80% of the total running time of a classical GA. This approach has been developed with the goal to present a new and efficient parallel scheme of the classical GA with better performances in terms of running time. In this paper, we will present a new asynchronous parallel Master/Slave scheme of the GA and will show the power of our approach with the classification extraction rules problem.