Optimization models and heuristic solution methods for the integrated fleet sizing and replenishment planning problem with candidate delivery patterns

In this thesis work, two mathematical models were developed to formulate two basic and extended versions the integrated fleet sizing and replenishment planning problem, and various solution methods were proposed. In the first section of the thesis, a model was created for replenishment planning and fleet size determination under pre-defined replenishment frequencies and vendor management policy. For solving the mentioned model, two different metaheuristic solution techniques, namely Approximate Dynamic Programming, and Problem Space Search, were proposed and tested on datasets inspired by real-life data. In the second section, a more generalized version of the model in the first section is considered, and a mathematical model is developed from a new perspective, with intuitive solution methods proposed. Replenishment patterns were considered instead of pre-defined recurring replenishment frequencies. Additionally, demand seasonality and vehicle renting options were taken into account. To solve the problem, three different versions of the Fix and Optimize heuristic were designed and tested on generated data. The quality of the results and solution times demonstrated the effectiveness of the proposed solution techniques.