Karima Belmabrouk

Graph Neural Networks (GNNs) have become a foundational tool for e-commerce recommendation systems, yet they consistently fail in zero-shot cold-start scenarios where new products enter the market without prior interactions. In this paper, we diagnose this failure as a structural vulnerability rather than a simple data sparsity issue. We introduce the concept of asymmetric learning, demonstrating that in severely imbalanced bipartite graphs, minority-type nodes (products) become disproportionately reliant on topological signals. By evaluating this phenomenon alongside a relatively balanced control dataset, we confirm that this performance collapse is a byproduct of the data structure rather than architectural design. To mitigate this limitation, we propose Similarity-Driven History Augmentation (SHA), a data-centric approach that assigns synthetic interaction histories to cold-start products by matching them with semantically similar established donors. To prevent these synthetic signals from degrading the representations of established nodes, we further introduce a decoupled hybrid framework alongside an enhanced SHA strategy that selectively filters active customers. Comprehensive evaluations across multiple real-world e-commerce datasets, including DataCo and Amazon Gift Cards, confirm the effectiveness and stability of our approach across different GNN architectures.

This chapter explores the convergence of sustainability, intelligent methods, and smart grids as a solution to the escalating energy demands and environmental concerns of the 21st century. It investigates how smart grids, empowered by machine learning, artificial intelligence, and other advanced techniques, can create a more efficient and environmentally friendly energy ecosystem. Key applications in load forecasting, demand management, renewable integration, and grid optimization are examined. The chapter acknowledges challenges like resource intensity and data privacy, advocating for robust regulations and responsible development to ensure equitable access to smart grid benefits. Ultimately, it presents a comprehensive analysis of this synergy, aiming to equip readers with the knowledge to navigate the complexities of the energy landscape and build a sustainable energy future.

This paper presents a novel approach to supply chain optimization using a matrix-based adaptation of Particle Swarm Optimization (PSO). We address the challenge of applying PSO, originally designed for continuous problems, to the discrete, multi-dimensional nature of supply chain systems. Our method evolves through two versions: Matrix-based Binary PSO (MBPSO) and its improved variant, Matrix-Probabilistic PSO (M-PPSO). Both versions represent solutions as matrices, allowing for complex interactions between supply chain entities. MBPSO uses binary values to indicate interactions but suffers from local optima issues. To overcome this limitation, we developed M-PPSO, which employs probabilities for a more nuanced representation. We evaluate these versions against traditional PSO and Binary PSO on benchmark problems, including the Rastrigin function and the Knapsack problem. Results show that M-PPSO significantly outperforms MBPSO, resolving local optima issues and achieving consistent optimal solutions within fewer iterations. M-PPSO also enables more realistic modeling by allowing buyers to interact with multiple sellers for the same product. This research contributes to the field by demonstrating the efficacy of matrix-based PSO in handling the complexities of supply chain optimization, offering a promising direction for future applications in discrete, multientity optimization problems.

This book chapter provides a concise yet comprehensive exploration of supply chain optimization through a multi-agent approach. Spanning historical foundations to contemporary intelligent methodologies, the literature review examines communication, collaboration, and coordination methods between agents. The analysis reveals achievements and challenges, emphasizing the pivotal role of intelligent agents in enhancing decision-making and adaptability. The chapter serves as a valuable resource for researchers and practitioners, offering insights into the dynamic landscape of supply chain management within the context of multi-agent systems. Through this exploration, the chapter contributes to shaping the future of resilient and efficient supply chain networks.

Sustainable energy planning and management is crucial for achieving a reliable, affordable, and sustainable energy supply, while promoting environmental protection, social equity, and economic efficiency. Optimizing renewable energy systems for sustainability involves the development, deployment, and operation of renewable energy systems that maximize their environmental, social, and economic benefits while minimizing negative impacts. Metaheuristic algorithms and computational modeling techniques are increasingly being used to optimize renewable energy systems for sustainability. This chapter highlights the use of computational modeling in the field of suitable energy planning and management, with a focus on some fields as solar power forecasting and wind turbine design. Two case studies are presented, namely the environmental economic power dispatch and the environmental friendly microgrids redundancy problem.

Cyber criminals pose a significant threat to data privacy, seeking to exploit weaknesses in digital security systems to gain access to sensitive information. Data privacy measures, such as antivirus software, firewalls, and employee training, are implemented to protect against unauthorized access and theft of personal and sensitive data. However, cyber criminals continue to find ways to circumvent these measures, leading to an ongoing arms race in the world of cybersecurity. To stay ahead of evolving threats, ongoing investments in research and development, as well as partnerships between government agencies, private sector organizations, and cybersecurity experts, are crucial in strengthening data privacy measures and protecting against cyber criminals. Request access from your librarian to read this chapter's full text.

This study proposes a computer‐assisted system with a design based on Polya's problem‐solving model. The system is designed to help average and low‐achieving second graders in mathematics with word‐based addition and subtraction questions. The emphasis of using the specific model was on dividing the problem‐solving procedure into stages and the concentration on the stages that are problematic for students. Specifically, we compared mathematical word problem‐solving performance and computational skills of students who utilized the computer‐assisted system with students who employed general strategy instruction. Participants consisted of 52 second‐grade students randomly assigned to treatment conditions. Students were pretested and posttested with mathematical problem‐solving and computation tests, and repeated measures of their progress with respect to word problem solving were registered. The results showed there was a significant difference between the experimental and control groups in terms of the word problem‐solving progress measure, favouring the experimental group. This confirms that providing students with a computer‐assisted system offered the opportunity to explore all stages of the problem‐solving procedure as one possible way to enhance their problem‐solving skills.

The evolution of the Internet and the competitiveness among companies were factors in the explosion of Web services. Web services are applications available on the Internet each performing a particular task. Web users often need to call different services to achieve a more complex task that can’t be satisfied by a simple service. And users often prefer to have the best services responding to their requests. In this context, we should measure the Quality of Service (QoS) which is a very important aspect of Web services in order to offer to the user the best services. “How can we ensure the composition of different services to respond the user request” is the first problem that we contribute to resolve, proposing a multi-agent based model for the automatic planification of Web services. And “guarantee the required quality of the composite Web services” is a complex task regarding the unpredictable nature and dynamics of composite Web services, so our contribution to remedy to this problem consists of the use of two classes of quality attributes. The first one considers generic and the second contains specific attributes.