Improvements to bio-inspired optimisers usually answer generic defects with generic operators. This paper instruments the red-billed blue magpie optimizer (RBMO) and measures what fails. Both operators scale their step by differences between population positions, and the attack places every candidate at the incumbent best, so the search scale shrinks with the spread the attack destroys. On five instrumented functions of the IEEE Congress on Evolutionary Computation 2017 (CEC2017) benchmark suite the population collapses to numerical identity, and on four the error stagnates over the last 28% to 58% of the run. The proposed collapse-resistant RBMO (CRBMO) breaks the loop with a dual-anchor attack, a radius control under the one-fifth success rule and an adaptive foraging allocation; it adds no chaotic map, Levy flight or opposition-based learning. On CEC2017 at 30,000 evaluations, against nine swarm algorithms, it attains the best Friedman rank at
Accurate recognition of individual appliances from their own power consumption profiles underpins non-intrusive load monitoring (NILM), yet deep learning models often incur high computational overhead and lack operational interpretability. This study proposes a new explainable and lightweight model named β-PruNet to recognize household appliances using daily power consumption signals. The task addressed here is appliance classification from submetered profiles rather than disaggregation of an aggregate household signal, so the model is positioned as the recognition stage of a NILM pipeline. Each 24 h signal (86,400 samples) is transformed into a 720-bin ordinal pattern (OP) distribution, and the distribution is summarized by 27 statistical and spectral descriptors over 48 sliding windows. A compact class-weighted convolutional neural network extracts 128 embedding features. Drawing inspiration from collective swarm intelligence and biological synaptic pruning, a particle swarm optimization (PSO)-tuned regularized extreme learning machine (R-ELM) classifier with beta-weight centrality pruning conducts classification via a diverse softmax ensemble. On an imbalanced 18-class benchmark containing 916 daily appliance profiles, β-PruNet achieved 98.9% classification accuracy and a 97.6% macro F1-score while pruning 87.3% of hidden neurons (from 1500 to 191) without accuracy loss. Post hoc explainability analyses using SHAP, LIME, and Grad-CAM confirmed that the pruning criterion strongly correlates with Shapley feature attribution (r = 0.974). These findings indicate that integrating symbolic dynamics with bio-inspired pruning delivers a computationally efficient, highly accurate, and transparent solution suitable for edge-oriented energy management systems. The theoretical contribution is that the magnitude of a hidden unit’s learned output weights provides a usable ranking signal at no additional computational cost beyond the closed-form solution already obtained. The practical contribution is that this signal allows the classifier head to be compressed by 87.3% without a measurable loss in accuracy.
Parasitic Architecture represents a radical strategy for urban densification, yet the literature lacks a systematic morphological framework to analyze its physical relationship with the host beyond metaphorical descriptions. Addressing this theoretical gap, this study proposes an action-based analytical typology to classify these interventions. Adopting a purposive, case-based qualitative content analysis, the research examines fifteen documented cases in order to trace how each intervention physically engages its host. The findings identify five distinct morphological attachment strategies: Wedging, Perching, Clinging, Piercing, and Wrapping. These actions are synthesized into a matrix that relates the host interface to the specific spatial strategy of the parasite. Consequently, the findings indicate that these forms can be organized according to their primary morphological action in relation to the host. This study ultimately repositions the concept of the parasite from a critical metaphor into an operational grammar of architectural intervention, offering an analytical tool with potential applications in adaptive reuse and urban design.
The coexistence of multiple chronic diseases is common among older adults and presents substantial challenges for disease prevention, management, and healthcare planning. However, most previous studies have focused on predicting individual diseases, although multiple chronic conditions commonly coexist. This study evaluates a multi-output classification model that predicts twelve chronic disease-status labels independently and examines genetic algorithm (GA)- and particle swarm optimization (PSO)-based feature selection using a shared feature subset. Data from the Korean Longitudinal Study of Aging (KLoSA) were analyzed (n=4529). Random forest, decision tree, and XGBoost models were evaluated with five repeated nested holdout splits so that feature selection did not use the outer test set. Across runs, GA and PSO reduced the 131-feature input to 49–77 features, but changes in Composite Macro-F1 were small and model-dependent. High mean label accuracy coexisted with low sensitivity for several rare disease-positive classes. SHAP analysis on held-out data was therefore interpreted descriptively and showed lower ranking stability for selected models. These results indicate a trade-off between feature reduction, predictive performance, and computational cost rather than a general performance advantage or evidence between disease-label interactions.
Effective constrained multiobjective optimization requires a search process to coordinate feasibility attainment, objective-space progress, and population distribution under a limited evaluation budget. This coordination becomes more difficult in dual-population frameworks when auxiliary search loses objective-direction information, applies unsuitable constraint tolerance, or consumes resources that no longer match its current contribution. Motivated by these limitations, we develop the Multi-Morphogen Field-Guided Constrained Multiobjective Evolutionary Algorithm (MFGCMOEA). Its main population is responsible for feasibility-oriented convergence, whereas the auxiliary population encodes each objective through a separate morphogen channel. Signals are propagated according to relative geometry in objective-rank space, allowing local search information to modify the developmental state of each candidate. These channel-wise states are then used to construct individual constraint buffers and multidimensional phenotypes, so promising infeasible directions can remain competitive without collapsing multiple objectives into a single scalar evaluation. Diversity regulation is applied only when truncation is required on the critical developmental front. For computational coordination, MFGCMOEA derives resource evidence from offspring developmental quality and realized auxiliary recruitment and combines it with a symmetric basal-growth term to prevent overly polarized offspring allocation. The method was evaluated on 47 benchmark problems from four suites using 30 independent runs. Among nine algorithms, MFGCMOEA obtained the best overall mean ranks, approximately 1.66 for IGD and 1.74 for HV. On the 26 problems with complete valid results for all methods, the Friedman test detected significant overall differences, while Holm-adjusted pairwise comparisons favored MFGCMOEA against all eight competitors with large rank-biserial effect sizes. Additional convergence, distribution, ablation, and diagnostic results further confirm the roles of the developmental-field representation and homeostatic resource regulation.
