Optimization problems are everywhere, in technology, economy, science, industry, environment, telecommunication, energy, transport, and supply chain management. They occur in private and public sectors, by individual or by small or huge enterprises. Although they are at first glance quite different, fortunately they usually share similar properties. In other words, a variety of problems in different areas of human activities share the similar symbolic presentation or mathematical model. They all have an objective function and the solution space of huge number of solutions, where just one or just a few among them are optimal.
The most useful and the most popular techniques in solving hard real-life optimization problems are so-called heuristic methods that provide approximate or near optimal solutions. Metaheuristics, or frameworks for building heuristic, are trying to establish general principles and rules that should be followed during the search for the better solution. Heuristic approach named ’Less is more approach’ (LIMA) has been recently proposed. Its main idea is to find the minimum number of search ingredients in solving some optimization problem that makes some heuristic more efficient than the currently best in the literature. More precisely, the goal is to make heuristic as simple as possible, but at the same time more effective and efficient than the current state-of-the-art heuristic. Several problems that follow LIMA idea have already been successfully implemented. They include area such as continuous optimization, dispersion, constrained clustering, etc.
In this project, we plan to continue research in LIMA direction and apply it to some new practical important problems:
In the field of Automatic Programming (AP), the solution of a problem is a program, which is usually represented by an AP tree. Application examples are symbolic regression, classification, and prediction that are very often used in engineering, energy planning, supply chains, etc.