Applications of X-Heuristics in Complex Combinatorial Optimization Problems pdf

In many industries, businesses need to make decisions in complex situations with multiple variables and constraints. These situations can often be formulated as combinatorial optimization problems (COPs). Solving these problems is essential to improve productivity and efficiency. However, this involves finding optimal solutions, which can be extremely challenging due to the NP-hard nature of some problems. This thesis proposes different x-heuristics to address the limitations of existing approaches for solving several COPs. These methodologies comprise different extensions of heuristic algorithms depending on the specific requirements of the problem. Specifically, biased-randomized heuristics combine heuristic algorithms with biased-randomization techniques to solve problems with deterministic parameters. Discrete-event heuristics combine heuristic algorithms with discrete-event simulation to address problems with synchronization issues, while agile optimization methods solve problems with real-time constraints. Simheuristics address problems with stochastic uncertainty, which combine heuristic algorithms with simulation. This is extended to fuzzy-simheuristics using fuzzy logic to solve problems with stochastic and non-stochastic uncertainty. In addition, this thesis proposes a novel methodology called sim-learnheuristics to solve problems with stochastic uncertainty and dynamic factors, which combines heuristic algorithms, simulation, and machine learning methods. The performance of these methodologies is validated through extensive computational experiments. The results highlight their efficiency in terms of solution quality and computational time, as well as their reliability in solving problems with multiple variables and constraints. Lastly, future research directions discuss real-time optimization methodologies for immediate decision-making capabilities in dynamic contexts.