Solving complex (multi-objective) optimization problems can be painstakingly difficult endeavor considering multiple and conflicting design goals. A growing trend in utilizing meta-heuristic algorithms to solve these problems has been observed as they have shown considerable success in dealing with tradeoffs between conflicting design goals. Many meta-heuristic algorithms have been developed in the past 25 years. Most of these algorithms do have merits, but they require tuning of their specified control parameters. For example, Genetic Algorithm require substantial tuning for population size, mutation and cross over rate. The same issue also appears in the case of Particle Swarm Optimization which relies on population size, inertia weight, social and cognitive parameters as parameters. In similar manner, Harmony Search requires tuning of harmony size, harmony memory consideration rate, and pitch adjustment. As for Ant Colony, the calibration of evaporation rate, pheromone influence, and heuristic influence are essential. In many cases, improper tuning for all of these specific parameters undesirably increases computational efforts as well as yields sub-optimal solutions. As a result, many researchers have advocated the adoption of parameter free meta-heuristic algorithms.
This tutorial introduces the basic of parameter free meta-heuristic algorithms. Additionally, this tutorial also reviews the current state-of-the-arts and introduces the new breed of parameter free meta-heuristic algorithms including Teaching Learning based Optimization, Sine Cosine Algorithm and Jaya Algorithm. The tutorial also includes hands-on exercise on adopting these algorithms along with comparative analysis of existing algorithm such as Particle Swarm Optimization and Cuckoo Search Algorithm and Flower Algorithm. Finally, the tutorial also introduces hybrid meta-heuristic algorithms based on the hyper-heuristics along with the technique to deal and deploy with multi-objective optimization problems.