Mapa Weba

Research Class: Optimization by Simulated Annealing: A Nature-Inspired Approach

Datum i mjesto održavanja: 27.11.2023., 11:00, O-357, u sklopu kolegija Meko Računarstvo
Predavač: dr. Tibor Lukic, full professor of theoretical and applied mathematics, Chair of Mathematics, Faculty of Technical Sciences, University of Novi Sad, Novi Sad, Serbia
Naziv predavanja: Optimization by Simulated Annealing: A Nature-Inspired Approach
Jezik predavanja: 
Engleski jezik/Hrvatski jezik

Sažetak predavanja:

Nature-inspired optimization algorithms are a set of problem-solving methodologies and approaches derived from natural processes. These algorithms are designed to solve complex optimization problems by mimicking the strategies used by various natural living systems to adapt, evolve, and survive in their environments or simulating methods in applied science or industry  based on natural laws. These algorithms can efficiently explore solution spaces and find near-optimal solutions to a wide range of problems.

Simulated Annealing (SA) is a well known stochastic optimization algorithm. The algorithm  simulates the process of steel hardening, which represents the slow cooling of the heated metal. In an annealing process, the observed system, initially at high temperature and high energy, is cooled by gradually lowering the temperature so that it is approximately in thermodynamic equilibrium at any time, until the system converges to a steady, ”frozen” ground state. The SA algorithm applies these steps, based on natural science laws, in order to find the optimal solution of the considered optimization problem.  The major advantage of the SA algorithm is its robustness: it does not require any derivative calculations of the objective function, which allows handling situations when the objective function has a complex form or even if it is non-differentiable. The presentation will show and analyze  a specific  application of the SA in the field of tomographic image reconstruction.

Životopis predavača

Podijeli članak