combinatorial multi armed bandit پیشرفته

نیاز به دانش پیشرفته کامپیوتر ساینس اول نوشتن الگوریتم به انگلیسی در فرم رسمی مقالات. دوم کد الگوریتم در پایتون و سیمولیشن سوم معیار های محاسبه الگوریتم و نمودار ها نوشتن تحلیل ها 2 صفحه based on algorithms of these papers: "Combinatorial Multi-Armed Bandit with General Reward Functions" "Contextual Combinatorial Multi-armed Bandits with Volatile Arms and Submodular Reward" "BUDGETED COMBINATORIAL MULTI-ARMED BANDITS" Advance level Computer science theory We need these specifically step by step: IMPORTANT TO READ CAREFULLY 1. Defining the environment : arms: each arm has a distribution which the agent doesn't know, and some characteristics. These are randomly selected. Rewards depend on the mean of the arm and a multiplier of how related the arms are. coherent matters. closer characteristics to each other the higher the reward. 2. Creating an Agent (algorithm) at each time stage, the agent selects a set of arms (max k number out of all N arms), then observes the rewards of the selected set. and the Agent should learn how are the characteristics of each arm, and at the end selecting the best combination of arms. 3. Training 4. test of Algorithm and seeing result. 5. Printing graphs After reading the papers carefully and understanding the problem we can start. Thanks Bests