Deadline Date: 25 May 2026
In the age of intelligent, data-centric engineering, the solution of inverse problems and optimizing designs effectively are two of the most challenging and consequential problems in engineering fields. Inverse problems for system parameters or latent inputs being recovered from observed outputs are generally ill-posed, nonlinear, and noise-sensitive, often with non-unique or unstable solutions. Simultaneously, engineering design optimization involves navigating high-dimensional, multi-objective, and often non-convex search spaces under uncertainty, which complicates the pursuit of optimal, reliable, and efficient designs. To cross these barriers, recent developments have moved increasingly towards hybrid intelligent strategies that combine the strengths of fuzzy logic, evolutionary algorithms, genetic programming, and swarm intelligence paradigms of soft computing with strong machine learning (ML) methods such as deep neural networks, reinforcement learning, Gaussian processes, and kernel-based models. The goal is to demonstrate how the combination of heuristic reasoning with data-driven learning can produce strong, interpretable, and scalable models that can manage uncertainty, take advantage of incomplete or noisy data, and achieve high performance on a range of complex engineering tasks.
Soft computing techniques are renowned for their ability to handle uncertainty, imprecision, and the vagueness inherent in complex real-world systems, offering heuristic approaches that mimic human reasoning. Meanwhile, machine learning (ML) excels in uncovering hidden patterns, enabling data-driven modeling, and delivering powerful predictive capabilities across diverse domains. The convergence of these two paradigms soft computing and machine learning offers a transformative approach to engineering problems, particularly in scenarios involving sparse, noisy, or incomplete data. This integration facilitates the creation of intelligent, adaptive, and robust computational frameworks that can efficiently learn from limited observations, adapt to dynamic environments and effectively navigate high-dimensional and nonlinear design spaces. Inverse problems such as material property estimation, medical image reconstruction, non-destructive evaluation, and geophysical modeling are often ill-posed and sensitive to noise, making them ideal candidates for hybrid intelligent solutions. Similarly, design optimization challenges in fields like structural and civil engineering, aerospace system design, fluid dynamics, robotics, renewable energy systems, and manufacturing require innovative algorithms that can manage multi-objective trade-offs, computational constraints, and uncertainty in system behavior.
We particularly encourage submissions that propose novel architectures, algorithms, and frameworks combining soft computing and ML, demonstrate rigorous evaluation using real-world datasets or simulations, and provide theoretical insights into convergence, stability, or interpretability. Works that employ surrogate modeling, physics-informed ML, uncertainty quantification, or transfer learning in inverse modeling or design contexts are also welcome. The issue aims to foster a cross-disciplinary dialogue between computational intelligence, applied mathematics, and engineering design communities.
Contributions are invited on, but not restricted to, the following themes:
- Hybrid Techniques Based on Chaos Optimization and Convolutional Neural Architectures in Engineering Systems
- Grey System-Based Machine Learning Models for Data-Scarce Design Optimization
- Quantum-Inspired Algorithms Coupled with Sparse Learning for Solving Ill-Conditioned Engineering Inverse Tasks
- Self-Adaptive Hybrid Architectures with Rough Set Theory and Ensemble Deep Learning
- Fusion of Bio-Inspired Computing and Kernel-Based Regression for Design Optimization Under Uncertainty
- Cognitive Optimization Strategies with Tabu Search and Transformer-Based Learning
- Hybrid Fuzzy Rule-Based and Reinforcement Policy Models for Inverse Dynamics Estimation
- Topology-Guided Inverse Solutions Using Deep Belief Networks and Evolution Strategies
- Hybrid Neuro-Evolution and Symbolic Regression for Engineering Parameter Identification
- Integrated Soft-Sensing Models Using Cultural Algorithms and Recurrent Neural Networks
- Adaptive Decision-Making in Inverse Design Using Artificial Immune Systems and Gradient Boosting
- Multi-Objective Optimization Through Harmony Search and Attention-Based Learning
- Hybrid Learning Frameworks with Stochastic Search and Capsule Neural Networks for Engineering Design Exploration
In the age of intelligent, data-centric engineering, the solution of inverse problems and optimizing designs effectively are two of the most challenging and consequential problems in engineering fields. Inverse problems for system parameters or latent inputs being recovered from observed outputs are generally ill-posed, nonlinear, and noise-sensitive, often with non-unique or unstable ... show more
Editorial board
