Unlike other block puzzle games, AI Simulator: Block Puzzle is a bit different. You can’t just drag and drop the blocks, you have to use AI to control where the blocks go.
You start out with a very simple random AI, which will randomly place the blocks.
You can unlock more advanced AIs using chips you earn from playing the game.
You can get chips by playing the game. You can also get double chips by watching ads at the end of each game.
By upgrading to Pro, you can get 2x chips for every game automatically without having to watch ads.
At the beginning, you should try to get CLLI (Clear Line Heuristic) in Basic Heuristics as soon as possible. It is the most basic AI that can solve the first few levels.
After that, you can try to get the other AIs such as HEUR (Heuristics) in Advanced Heuristics. These algorithms combine different heuristics together and are more powerful than algorithms in Basic Heuristics.
Algorithms such as MOCA (Monte Carlo Simulation) use Monte Carlo Methods to simulate many possible moves and choose the best move. These algorithms are quite powerful but very slow.
Eventually, you can unlock machine learning algorithms such as DQN (Deep Q Learning) and PPO (Proximal Policy Optimization, currently Work-In-Progress). These algorithms are very powerful, but they need a lot of training to get good results.
It is good to use the default parameters for machine learning algorithms if you are just starting out. After you get the hang of it, you can try to tune the parameters to get better results.
You can learn more about the DQN algorithm here.