beyond norms

Innovation Science and Engineering Research

Quantum

Intelligence

Optimization

Radios

Cybersecurity

 © 2024 axon logic. All Rights Reserved

    Offices: 1 Thriamvou, 141 22, Athens, Greece

    Labs: 21 M. Timotheou, 142 31, Athens, Greece

    Reg. No.: 326575 (Chamber of Commerce & Industry Athens)

    +30 210 2833 116  ||  + 30 6944 6962 21  || 

Networking

Engineering is all about optimization (the process of doing the most with the least).

For example, upon designing an Industrial Internet of Things scheduler, network topology and transmit power performance must be optimized for minimal Latency response and Energy consumption, while keeping maximal service reliability and network coverage.

Or, the design of a typical machine-learning problem, involves optimization at multiple steps in addition to the learning model, including the choice of the hyperparameters of the model, the transforms to apply to the data prior to modeling, the modeling pipeline to use as the final model, and so on.

In this respect, we put much effort in developing our own optimization objective functions and solution methods that can make systems as efficient as possible by finding the most effective or favorable trade-offs between performance(s) or condition(s).

Particularly, we are specialized in formulating multi-objective and multi-modal optimization objective functions and problems that can improve sets of multi-dimensional trade-off designs upon numerous criteria - this can better articulate (and thus capture) the demands of complex systems in comparison to traditional one-dimensional techniques, where single-function variables and constraint criteria are typically considered. 

We further use our own developed convex optimization-based solution methods that are non-programming and independent from duality, meaning that we can guarantee convergence with the minimum implementation complexity of the final optimization algorithm.

These methods also include specialized mathematical analysis and tricks to avoid recursiveness (a state that often occurs in optimization processes), such that our algorithms can approximate transcendental equations with the minimum truncation error and implement in close-to-real-time fashion.

Bio-inspired optimization is also of interest. We are experienced in creating distributed decision-making processes for telecommunications systems, using principles of biological systems, such as particle swarm optimization, ant-bee-bat-firefly algorithms, Cuckoo and Harmony searches, simulated annealing, differential evolution, and Genetic algorithms.

The development of our optimization algorithms is tailored to the given problem and the final optimized performance(s) can be adjusted according to the application.

Customers can inquire more details here.