MIS41480

The optimization problems are everywhere in our daily lives. Whether it's finding the best house we can afford, minimizing energy consumption at home, taking the shortest path to our destination, investing in the most profitable portfolio, or creating a schedule that meets all our deadlines, we often rely on our own intuitive methods. In this module, we will explore more structured, scientifically grounded approaches to these intuitive methods, known as “heuristics”.

Heuristic algorithms are problem-solving methods designed to find good (but not necessarily the best) solutions to complex problems, which are often inspired by experience or intuition. In this module, we will first examine a variety of optimization problems and discuss their similarities, challenging the assumption they are fundamentally different. We will then explore prominent exact approaches used to find the best solutions (e.g., branch and bound algorithm and dynamic programming) and discuss when heuristic algorithms may offer a better, more practical alternative. .

We will study the fundamentals of heuristic algorithms and identify the common attributes of efficient heuristic algorithms (e.g., local search and diversification). Metaheuristics, that are high-level frameworks designed to guide the development of heuristic algorithms, will be introduced as a way to leverage these common attributes. We will cover several well-known metaheuristic algorithms (e.g., simulated annealing, tabu search and genetic algorithms).

We will also address various versions of heuristic algorithms (e.g., multi-objective heuristics, matheuristics which combine exact and heuristic methods). Finally, we will examine different applications of heuristic algorithms in real-world problem-solving scenarios.