Słownik AI
Kompletny słownik sztucznej inteligencji
Quantum Particle Swarm Optimization (QPSO)
A variant of Particle Swarm Optimization that incorporates principles of quantum mechanics, particularly state superposition and the delta potential, to improve convergence and avoid local optima.
Quantum Potential Well
A mathematical model in QPSO where each particle is confined within a search space defined by a potential well, whose position and width are dynamic and influence the particle's movement.
Attractor Position (pbest)
In QPSO, the best position found by a particle so far, which acts as a center of attraction in its quantum potential well, unlike the velocity concept in classical PSO.
Center of Mass (Mean Best Position)
The average position of the personal best positions (pbest) of all particles in the swarm, serving as a global reference point for calculating the quantum movement of each particle.
Contraction-Expansion Parameter (CEC)
A unique coefficient in QPSO that controls the balance between exploration and exploitation by adjusting the size of the quantum potential well, replacing the inertia, cognitive, and social parameters of classical PSO.
Position Probability Distribution
A probability density function (often an exponential or Cauchy distribution) describing the probability of a particle being at a certain position in its potential well, determining its next location.
Quantum State Superposition
A principle applied in QPSO where a particle does not have a defined position but exists in a probability distribution over multiple positions simultaneously, allowing for broader exploration of the search space.
Characteristic Distance (L)
A variable in QPSO that defines the length of the potential well for each particle, calculated from the distance between its current position and the swarm's center of mass, directly influencing its movement amplitude.
Quantum Evolution Equation
Fundamental mathematical formula of QPSO that updates a particle's position using a random variable drawn from a probability distribution, the center of mass, and its best personal position.
Quantum Escape
Phenomenon in QPSO where a particle can 'escape' from a local optimum thanks to the probabilistic nature of its movement, simulating the quantum tunneling effect to improve global search capability.
QPSO-Classical Hybridization
Approach combining the update mechanisms of traditional PSO (velocity-based) with the potential well model of QPSO to leverage the exploration and exploitation strengths of both methods.
Multi-swarm QPSO
Variant of QPSO where multiple sub-swarms evolve in parallel, each with its own center of mass, to increase search diversity and reduce the risk of premature convergence to a suboptimal solution.
QPSO Convergence Bound
Theoretical analysis guaranteeing that, under certain conditions on the CEC parameter, the QPSO algorithm will converge to a fixed point, ensuring the stability and reliability of the optimization process.
Quantum Neighborhood
Topological structure in QPSO where a particle's influence on its neighbors is defined probabilistically rather than by a fixed distance, allowing for dynamic and adaptive interactions within the swarm.
Adaptive QPSO
Version of QPSO where the contraction-expansion parameter (CEC) is dynamically adjusted during optimization based on performance metrics such as swarm diversity or the solution's improvement rate.
Dirac Delta Potential
Simplified potential model used in some QPSO implementations, where the potential well is represented by a delta function, leading to a specific position distribution and a simplified update equation.