Jorge Mario Cruz-Duarte

This work introduces an alternative approach that utilizes a Hyper-Heuristic framework under the umbrella of Automated Algorithm Design to dynamically tailor Metaheuristics (MHs) for fine-tuning PID controller parameters. Our findings emphasize the robust consistency, high precision, and adaptability of the method in creating custom, metaphor-free MHs. We employ this methodology in a case study of real-world engineering problems to develop a control-heuristic-based tuner that significantly improves the performance of DC-DC Buck converters across diverse operational conditions. The analysis reveals dominant trends in search operator preferences, highlighting the reliance on genetic crossover and swarm dynamics. These insights provide actionable directions for similar applications in power electronics. Through rigorous statistical comparisons, we verify the superior efficacy and reliability of our approach relative to conventional methods, remarking on its potential for broad application in the field.

In targeted adversarial attacks on vision models, the selection of the target label is a critical yet often overlooked determinant of attack success. This target label corresponds to the class that the attacker aims to force the model to predict. Now, existing strategies typically rely on randomness, model predictions, or static semantic resources, limiting interpretability, reproducibility, or flexibility. This paper then proposes a semantics-guided framework for adversarial target selection using the cross-modal knowledge transfer from pretrained language and vision-language models. We evaluate several state-of-the-art models (BERT, TinyLLAMA, and CLIP) as similarity sources to select the most and least semantically related labels with respect to the ground truth, forming best- and worst-case adversarial scenarios. Our experiments on three vision models and five attack methods reveal that these models consistently render practical adversarial targets and surpass static lexical databases, such as WordNet, particularly for distant class relationships. We also observe that static testing of target labels offers a preliminary assessment of the effectiveness of similarity sources, \textit{a priori} testing. Our results corroborate the suitability of pretrained models for constructing interpretable, standardized, and scalable adversarial benchmarks across architectures and datasets.

The increasing energy footprint of artificial intelligence systems urges alternative computational models that are both efficient and scalable. Neuromorphic Computing (NC) addresses this challenge by empowering event-driven algorithms that operate with minimal power requirements through biologically inspired spiking dynamics. We present the NeurOptimiser, a fully spike-based optimisation framework that materialises the neuromorphic-based metaheuristic paradigm through a decentralised NC system. The proposed approach comprises a population of Neuromorphic Heuristic Units (NHUs), each combining spiking neuron dynamics with spike-triggered perturbation heuristics to evolve candidate solutions asynchronously. The NeurOptimiser's coordination arises through native spiking mechanisms that support activity propagation, local information sharing, and global state updates without external orchestration. We implement this framework on Intel's Lava platform, targeting the Loihi 2 chip, and evaluate it on the noiseless BBOB suite up to 40 dimensions. We deploy several NeurOptimisers using different configurations, mainly considering dynamic systems such as linear and Izhikevich models for spiking neural dynamics, and fixed and Differential Evolution mutation rules for spike-triggered heuristics. Although these configurations are implemented as a proof of concept, we document and outline further extensions and improvements to the framework implementation. Results show that the proposed approach exhibits structured population dynamics, consistent convergence, and milliwatt-level power feasibility. They also position spike-native MHs as a viable path toward real-time, low-energy, and decentralised optimisation.

Scheduling problems, which involve allocating resources to tasks over specified time periods to optimize objectives, are crucial in various fields. This work presents hyper-heuristic applications for scheduling problems, analyzing 215 peer-reviewed publications over the last decade. We categorize and examine the prevailing strategies and configurations of hyper-heuristics, mainly focusing on their application across diverse scheduling scenarios such as job shop, flow shop, timetabling, and project scheduling. Our findings reveal a strong inclination towards selection and perturbative hyper-heuristics, with evolutionary computation emerging as the most commonly employed technique in this context, particularly in job shop and timetabling problems. Despite the robust development in hyper-heuristic methodologies, our analysis indicates an under-representation of multi-objective optimization and a limited use of performance metrics beyond makespan and tardiness. We also identify potential areas for future research, such as expanding hyper-heuristic applications to underexplored industries and exploring less conventional performance metrics. By providing a comprehensive overview of the current landscape and outlining future research directions, we aim to guide and inspire ongoing innovations in scheduling problem research.