The Improvemence And Application Of Teaching And Learning Optimization Algorithm

Teaching Learning based Optimization Algorithms( TLBO) is a novel swarm intelligence algorithm proposed in recent years to simulate the real-life teaching and learning process in order to improve academic performance, the algorithm has features like:simplicity, scalability, flexibility, robustness, self-organization, implicit parallelism, etc.And it's widely used in many fields. Research shows that, TLBO algorithm is still has a slow convergence, low precision to solve problems, easy to fall into local optimum weaknesses. In order to improve the efficiency of optimization algorithms and solve deficiencies existed, this work proposes two improved mechanisms: first multi-species are based on the mechanism of teachers learning from each other(More Teaching Learning based Optimization Algorithms, MTLBO), the second is based on the adaptive step(Adaptive Teaching Learning based Optimization Algorithms, ATLBO). The main work is as follows:(1) The application of stochastic process theory analyzed TLBO convergence of the algorithm.(2) For solving the problems like late slow convergence, low precision, easy to fall into local optimum weakness of the standard TLBO, an improved mechanism for more teachers learn from each other are advanced.Setting the number of teachers teaching in TLBO algorithm to maintain the diversity of the population. And among teachers can also learn from each other in order to improve students' learning speed, thereby improving the accuracy of the optimization algorithm and overcoming the weakness of falling into local optimum. The improved algorithm on 10 benchmark test functions with GA, PSO, AFSA and the standard TLBO comparison, test results show that the improved algorithm has a faster convergence speed and higher precision solution.(3) In order to overcome the standard algorithm TLBO low precision, easy to fall into local optimum weakness, an adaptive step improvement mechanism. Standard algorithms,students learn step is a random value, ignoring the actual rate of progress with high school students and the quality of its state to change the law. Improved learning step with the change in the students' own status is changed, thereby improving the accuracy of the optimization algorithm. By testing on 10 Benchmark function with GA, PSO, AFSA andthe standard TLBO comparison, test results show that the improved algorithm in the solution accuracy, stability and convergence speed are better than TLBO.(4) The algorithm is applied to solve the problem of virtual logistics, and has achieved good results.