A novel computational procedure based on league championship algorithm for solving an inverse heat conduction problem

Abstract

Inverse heat conduction problems, which are one of the most important groups of problems, are often ill-posed and complicated problems, and their optimization process has lots of local extrema. This paper provides a novel computational procedure based on finite differences method and league championship algorithm to solve a one-dimensional inverse heat conduction problem. At the beginning, we use the Crank-Nicolson semi-implicit finite difference scheme to discretize the problem domain and solve the direct problem which is a second-order method in time and unconditionally stable. The consistency, stability and convergence of the method are investigated. Then we employ a new optimization method known as league championship algorithm to estimate the unknown boundary condition from some measured temperature on the line. League championship algorithm is a recently proposed probabilistic algorithm for optimization in continuous environments, which tries to simulate a championship environment wherein several teams with different abilities play in an artificial league for several weeks or iterations. To confirm the efficiency and accuracy of the proposed approach, we give some examples for the engineering applications. Results show an excellent agreement between the solution of the proposed numerical algorithm and the exact solution.