Modelling time complexity of micro-genetic algorithms for online traffic control decisions

Abstract

This paper describes an experimental approach to test the suitability of micro-genetic algorithms (m-GAs) to solve large combinatorial traffic control problems and establishes relationships between time to convergence and problem size. A discrete time dynamical traffic control problem with different sizes and levels of complexity was used as a test-bed. Results showed that m-GAs can tackle computationally demanding problems. Upon appropriately sizing the m-GA population, the m-GA converged to a near-optimal solution in a number of generations equal to the string length. Results also demonstrated that with the selection of appropriate number of generations, it is possible to get most of the worth of the theoretically optimal solution but with only a fraction of the computation cost. Results showed that as the size of the optimisation problem grew exponentially, the time requirements of m-GA grew only linearly thus making m-GAs especially suited for optimising large-scale and combinatorial problems for online optimisation.