An optimization algorithm "Team Model"
Tetsuyuki Takahama, Setsuko Sakai · 2003
There are many optimization problems where it is very difficult to obtain the exact solution within a practical time. In this case, we use the approximation algorithms, such as genetic algorithm (GA). We propose a new approximation algorithm "Team Model", which models teaching by teams. When an organization tries to improve the productivity of the performance as a whole, it needs a lot of materials and facilities if all individuals are educated at once. So, usually it forms a team of some individuals and educates the individuals in the team. We can apply this idea to optimization. In Team Model, one selects some individuals in the organization regardless of the fitness, forms the team with the individuals, executes the teaching operation according to the fitness, and obtains better individuals. In GA, on the contrary, one selects individuals according to the fitness, executes the operation such as crossover regardless of the fitness. We show the ability of Team Model by comparing it with GA through computer simulation of some benchmark problems for optimization.