This article discusses an algorithm portfolio approach to find optimal set-up plans in a dynamic shop floor environment where flexibility and promptness of the decision process is critical along with best possible utilisation of the available resources. An evolutionary algorithm based reconfigurable set-up planning approach is presented where the final set-up plan is determined in two steps: primitive set-up planning through feature grouping and reconfigurable set-up merging based on real time information from the scheduling system. The tendency of single algorithm approach to converge to sub-optimal solutions was countered by using portfolios of genetic algorithm and its three variants: Genetic Algorithm with Chromosome Differentiation, Sexual Genetic Algorithm and a modified version of Age Genetic Algorithm. Best performing portfolios selected after exhaustive experimentation showed dramatic computational improvements in achieving the optimal solution validating the appropriateness and effectiveness of algorithm portfolio approach.