Teaching-Learning Based Optimization Algorithm Based on Course by Course Improvement

Teaching-learning-based optimization (TLBO) algorithm is a new intelligence algorithm. It takes evaluation scheme which updates learners´ grade totally in solving multi-dimensional function optimization problems, thus decreases the convergence speed. Moreover, the interference phenomena among dimensions would influence the quality of solutions. While all existing improvement strategies on TLBO introduce new parameters, thus destroy the simplicity of TLBO. To solve the above problems, a TLBO algorithm based on Course by course Improvement (TLBOCI) was proposed. This algorithm takes strategy which updates learners´ grade based on course by course improvement in teaching and learning phase in each iteration, and this strategy combines an updated value of one course with the values of other courses into a new solution. The results of four classic test functions show that the TLBOCI algorithm could outperform TLBO in quality of solutions and convergence speed under the condition of not introducing new parameters thus keeping the simplicity of TLBO.