Number Grid Coursework

Digital Grid Course My job is to study a box of 100 square 2 x 2 I will take a 2 x 2 square in a 100 square grid and multiply two angles. Then we will study the relationship between the two results by finding out the difference. Test 1 ====== 54 55 54 x 65 = 3510 64 65 55 x 64 = 3520 3520 - 35 10 = 10 difference = 10 test 2 ====== 5 6 5 x 16 = 80 15 16 6 x 15 = 90 90-80 = 10 difference = 10 test 3 ====== 18 19 18 x 29 = 522 28 29 19 x 28 = 532 532 - 52 = 10 difference = 10 prediction ========= = We predict that the difference between squares of 2 and 2 is always 10. =====

: Pixels (picture elements) are images that can be represented as digital grids. Each small square in the grid is a pixel. One or more values ​​can be associated with pixels, each of which typically represents intensity of color. Since the different colors consist only of red, green and blue, the most common pixel format consists of three numbers, each representing color intensity of each color.

If you are familiar with RGB, you will notice that almost all modern display technologies are made up of grids of red, green, blue pixels (this grid is called subpixel, It helps to display color). RGB is an additive color mixture. In other words, all the values ​​of red, green, and blue are built in brightness values, and if you combine various variations of red, green and blue according to the brightness, the colors will be different. The RGB format does not require a conversion process or upgrade, so it is the best way to transfer the data displayed on the RGB panel. The only disadvantage is that RGB is not compressed and the signal stream at 1920 x 1080 @ 60 hz is about 9 Gbps (not including alpha / transparent channels), which is too high for modern fiber optic internet connections.

The Grid World implementation aims to start the proxy from the upper left corner of the (0,0) grid and reach the lower right corner of the (Ny - 1, Nx - 1) grid angle with minimal steps. Ny + Nx step agent allows only one grid square move up, down, left, and right direction. In order to defeat this game, we used Policy Enhancement Monte Carlo Rewarded Average Sampling Reinforcement Learning Algorithm and used epsilon greedy proxy to navigate the grid world environment. Rewards, Status - In this implementation tabular action values ​​and policies are used (this applies to small grid sizes), and most of the above theory is in Sutton & Barto's "Reinforcement Learning" textbook. The Python code implementation of this algorithm is in Github.