The ever-increasing complexity and difficulty of optimization problems that humankind faces have substantially increased the demand for efficient and accurate optimization algorithms. Over the last two decades, nature-inspired stochastic optimization techniques have been widely used to solve a variety of problems in both science and industry. They have indeed become reliable alternatives to conventional optimization algorithms. Despite the popularity and substantial impact of such methods, optimization of real-world problems involves addressing a large number of difficulties, including but not limited to locally optimal solutions, multiple objectives, constraints, expensive objective functions, noisy data, dynamic search spaces, and uncertainties. In this talk, such difficulties are first presented along with their impact on the whole optimization process. The idea of agile and iterative optimization workflows between optimization experts and decision makers is then discussed to ensure effective use of optimization algorithms and minimize the risk of failure. Challenges and recent advances in the design and development of exact, heuristic, meta-heuristic, and hyper-heuristic optimization algorithms are then covered. There will be also information on the reason why most recent optimization algorithms tend to be nature-inspired, the future of research in this interesting area, and the substantial impact on societies.