Mohlo by se vám také líbit

Více informací o knize Combined Use of Reinforcement Learning and Simulated Annealing

• Parametry knihy
• Anotace
• Oblíbené z jiného soudku

Nákupem získáte 149 bodů

Anotace knihy

In the dissertation combined reinforcement learning§(RL) and §simulated annealing (SA) concepts, problems, proposed§solutions, §algorithms and application examples are shown.§RL models a decision maker as a goal-driven agent§aiming to reach §goal states in the problem representation state§space. The agent §takes different choices among the numerous§possibilities, but each §choice can make different impact in the environment.§Each decision §has some effect being expressed in the form of§numeric honor or §dishonor, in a reward value. The agent utilizes the§feedback to §recognize which actions are honored and which are§not. The agent §then tries to govern its decision sequence into the§direction that §maximizes the environment s satisfaction .§The concept of SA is based on the analogy of how§liquids freeze. §There an initially high temperature and disordered§melt is slowly §cooled down and reaches thermal equilibrium.§While in annealing the temperature parameter bounds are §straightforward, in SA they might be dependent on the§problem and §its numeric representation.§This dissertation gives a method which can be used§for defining §temperature bounds in RL environment.

Parametry knihy

• Plný název: Combined Use of Reinforcement Learning and Simulated Annealing
• Podnázev: Algorithms and Applications
• Autor: Peter Stefan
• Jazyk: Angličtina
• Vazba: Brožovaná
• Počet stran: 128
• EAN: 9783639125931
• ID: 06821373
• Nakladatelství: VDM Verlag Dr. Müller
• Rozměry: 219 × 149 × 13 mm
• Rok vydání: 2009

Oblíbené z jiného soudku

---

---

---