Abstract This paper explores an automated approach to tuning constraint solvers using genetic algorithms (GAs). In traditional constraint programming, selecting preprocessing parameters is a manual process that requires expertise, creating a barrier for novices. The authors propose a self-learning genetic algorithm (SLGA) that leverages knowledge from minor problem instances to guide the search for optimal preprocessing in larger instances. SLGA begins by solving small-scale constraint satisfaction... This paper explores an automated approach to tuning constraint solvers using genetic algorithms (GAs). In traditional constraint programming, selecting preprocessing parameters is a manual process that requires expertise, creating a barrier for novices. The authors propose a self-learning genetic algorithm (SLGA) that leverages knowledge from minor problem instances to guide the search for optimal preprocessing in larger instances. SLGA begins by solving small-scale constraint satisfaction problems to identify effective preprocessing strategies. These strategies then form the starting population for the GA when applied to larger problems, replacing the standard random initialization. Two strategies are tested: Learning from Best (LFB) and Learning from Genetic (LFG), with LFB utilising the best-known methods and LFG employing those derived from previous GA runs. Experiments on benchmark problems (BIBD, N-Queen, Golomb, and Langford's number) show that SLGA outperforms standard GAs in both efficiency and solution quality. Particularly, LFG consistently finds better preprocessing settings, even when exhaustive search is infeasible. The paper concludes that SLGA is a promising tool for automating configuration in constraint solving, and future work will explore its application to more complex and large-scale problems. Read More Read Less

Abstract This paper presents a compositional and algebraic model for characterizing software and hardware based on their resource usage, with a focus on energy consumption and completion time. Recognising the limitations of instruction-level profiling and the nondeterminism introduced by modern system complexities, the authors propose a black-box approach that measures the overall resource usage of programs without requiring source code access. By treating programs as... This paper presents a compositional and algebraic model for characterizing software and hardware based on their resource usage, with a focus on energy consumption and completion time. Recognising the limitations of instruction-level profiling and the nondeterminism introduced by modern system complexities, the authors propose a black-box approach that measures the overall resource usage of programs without requiring source code access. By treating programs as combinations of computational patterns, the model uses synthetic benchmarks as building blocks to represent programs in terms of their resource behavior. The authors introduce a linear algebra framework where resource usage data is captured in matrices, enabling decomposition of programs into benchmark-based components. The model defines three conceptual spaces—measurement space, splitup space, and benchmark space—to analyze program behavior, system characteristics, and resource efficiency. This enables comparison of software across systems and identification of unknown computational patterns. Experimental results on a desktop system demonstrate how the model captures the growing dominance of memory operations in mergesort as input size increases. The approach offers a scalable, source-independent method for profiling and predicting energy and performance behavior across heterogeneous environments. Future work will explore applications in cloud computing, virtual machines, and heterogeneous architectures for real-time resource forecasting. Read More Read Less

Abstract Compressive Sampling (CS) is a new method of signal acquisition and reconstruction from frequency data which do not follow the basic principle of the Nyquist-Shannon sampling theory. This new method allows reconstruction of the signal from substantially fewer measurements than those required by conventional sampling methods. We present and discuss a new, swarm based, technique for representing and reconstructing signals, with real values, in... Compressive Sampling (CS) is a new method of signal acquisition and reconstruction from frequency data which do not follow the basic principle of the Nyquist-Shannon sampling theory. This new method allows reconstruction of the signal from substantially fewer measurements than those required by conventional sampling methods. We present and discuss a new, swarm based, technique for representing and reconstructing signals, with real values, in a noiseless environment. The method consists of finding an approximation of the l0-norm based problem, as a combinatorial optimization problem for signal reconstruction. We also present and discuss some experimental results which compare the accuracy and the running time of our heuristic to the IHT and IRLS methods. Read More Read Less