The linear mixed model constrained particle swarm optimization for hyperspectral endmember extraction from highly mixed data

2016 ◽  
Author(s):  
Mingming Xu ◽  
Liangpei Zhang ◽  
Bo Du ◽  
Lefei Zhang
Keyword(s):  
Particle Swarm Optimization ◽  
Mixed Model ◽  
Linear Mixed Model ◽  
Particle Swarm ◽  
Mixed Data ◽  
Endmember Extraction ◽  
Swarm Optimization

PSONK: PARTICLE SWARM OPTIMIZATION WITH NEGATIVE KNOWLEDGE FOR MULTI-OBJECTIVE U-SHAPED ASSEMBLY LINES BALANCING WITH PARALLEL WORKSTATIONS

2013 ◽  
Vol 12 (01) ◽  
pp. 15-41 ◽  
Author(s):  
PARAMES CHUTIMA ◽  
SUCHADA KID-ARN
Keyword(s):  
Particle Swarm Optimization ◽  
Mixed Model ◽  
Particle Swarm ◽  
Assembly Line Balancing ◽  
Pso Algorithm ◽  
Swarm Optimization ◽  
Multi Objective ◽  
Negative Knowledge ◽  
Conflicting Objectives ◽  
Mixed Model Assembly Line

Particle swarm optimization (PSO) is a metaheuristic method inspired by the swarming behavior observed in flocks of birds. This paper proposes a novel PSO method, namely the PSO algorithm with negative knowledge (PSONK), to solve the multi-objective mixed-model assembly line balancing (MULB) problem with parallel workstations. PSONK employs the knowledge of relative positions of different particles as opposed to traditional PSO in generating new solution strings. The knowledge about poor solutions is also utilized to avoid the pairs of adjacent tasks appeared in the poor solutions from being selected as parts of the new solution strings in the next generation. The concept of Pareto optimality is employed to allow the conflicting objectives to be optimized simultaneously. Experimental results show clearly that PSONK is a promising algorithm. In addition, PSONK embedded with an appropriate local search (M-PSONK) can result in improved performances.

Sign in / Sign up

Export Citation Format

Share Document