<p>There are several optimization techniques available for solving Metaheuristics optimization problems. Many of those required manual parameters tuning to get optimum results such as Crossover Probability (Pc), Mutation Probability (Pm), Population Size (N) in Genetic Algorithm (GA); Inertia Weight (w), Cognitive Coefficient (c₁) and Social Coefficient (c₂) in Particle Swarm Optimization (PSO). This additional tuning increases the computational complexity and space complexity for high dimensional optimization problems. The wrong choice of these values may result in ineffective results or prolonged execution of the computation. To solve this problem, the researchers from optimization community proposed the concept of parameter-less optimization algorithms in which we do not require user-defined parameters or those parameters are automatically adjusted during their runtime. This paper aims to provide a brief review of these algorithms including the core concept and the working principles. The discussion on techniques like Teaching-Learning-Based Optimization (TLBO), Jaya algorithm and Rao algorithms, Flood Algorithm, Best–Worst–Random (BWR) and Best–Mean–Random (BMR) give the idea of parameter-less algorithms to test benchmark functions which are widely used to check performance of any optimization algorithms such as Congress on Evolutionary Computation (CEC). These algorithms proved their superiority on unimodal, multimodal, hybrid, and multi-objective, real life design case studies, and composition functions. Statistical Tests and Convergence curves suggest that parameter-less algorithms have ability to converge at faster rate which increases their popularity in hyper-parameters optimization in deep learning, Image Processing and Video Processing. This study proposes the roadmap to develop the parameter-less algorithm based of literature.</p>