Population-based optimization algorithms start with initializing a random population and iteratively move towards the optimal solution. Most of these algorithms use the expression, “Lower Bound + random*(Upper Bound – Lower Bound)” to initialize the random solutions. This paper analyses four initialization strategies; the widely used one, the strategy used by “Jellfish Optimizer” and two newly proposed strategies. Among the newly proposed strategies, the first one is a modified version of the strategy used in the “Jelly Fish Optimizer” and the other one is based on a larger initial population combined with the concept of “Opposite Number”. Fifty-five benchmark functions including real-world industrial problems are considered in this work for the performance analyses. The obtained results demonstrate the better performance of the larger initial population-based strategy that uses the “Opposite Number” concept in both constrained and unconstrained benchmark instances. For fair analysis, the number of function evaluations and population size are kept the same.