MapReduce teaching learning based optimization algorithm for solving CEC-2013 LSGO benchmark Testsuit

Abstract

Teaching Learning Based Optimization (TLBO) algorithm, introduced in 2011 is widely used in optimization problems across various domains. It is a powerful tool that is capable of solving complex, multidimensional, linear, and nonlinear problems. MapReduce is a distributed programming model developed by Google. It is widely used for processing large datasets in parallel way. This paper proposes the use of the MapReduce programming paradigm for the implementation of the TLBO algorithm on distributed systems, creating a novel approach known as MapReduce Teaching Learning Based Optimization (MRTLBO). The proposed MRTLBO algorithm is tested on Congress of Evolutionary Computations (CEC)-2013 Large-Scale Global Optimization Benchmark Problems dataset, and its performance is compared with sequential TLBO algorithm on the same dataset. The experimental output exhibits that the MRTLBO algorithm is effective in working with high-dimensional problems, and it outperforms the sequential TLBO algorithm with respect to the final result, and speedup. Overall, the proposed MRTLBO algorithm gives a scalable and effective optimization strategy for working on optimization problems in distributed systems.