Reinforcement-Learning-with-MATLAB

This repository contains series of modules to get started with Reinforcement Learning with MATLAB.

SO_MCDM_SupplierSelection

As a multi-criteria decision-making (MCDM) problem, supplier selection plays a key role in achieving the objectives of a supply chain system. Multiple strategic, operational, quantitative, and qualitative criteria influence the supplier selection process. A wide spectrum of criteria have been introduced, classified and used by researchers and practitioners to evaluate the suppliers’ performance; however, measuring and employing all of these criteria is impractical in real-world scenarios due to the budget, time, and information limitations. In this study, a decision support system (DSS) is developed, which helps managers to select a set of most effective criteria for the supplier selection process. This DSS is a threefold integration of MCDM and simulation and optimization. In this framework, the MCDM module incorporates a combination of criteria to select the suppliers. Then, a simulation model is used to evaluate the performance of the supply chain system considering the selected suppliers. Based on the simulation results, a multi-objective metaheuristic algorithm is utilized to find the ideal combinations of the criteria to maximize the supply chain system performance.

MO-PASS

Appointment scheduling (AS) is one of the key factors used to improve patient satisfaction in healthcare services. A practical and robust appointment scheduling solution allows clinics to utilize medical assets, equipment, and resources in an efficient manner. This study introduces a Multi-Objective Patient Appointment Scheduling System (MO-PASS) to enhance clinic operations and patients’ satisfaction. This framework takes advantage of Multi-objective Particle Swarm Optimization (MOPSO) to optimize objectives simultaneously without negatively affecting others. To address stochastic parameters and the uncertain nature of the AS problem, this algorithm is used in a simulation-optimization setting to create a simheuristic model. The efficiency of the proposed framework is tested in a breast cancer clinic system with multiple physicians and patient types. Finally, the MO-PASS performance is compared against three heuristic approaches and its results were promising.

RL_Workshop_Series

RL_labs

This repo includes all required labs for Reinforcement Learning class taught by Prof. Dehghani