mimocha / ga-logic-circuit

Generation of logical circuit with genetics algorithm and cellular automata. Dissertation project for BEng Innovative Systems Engineering at Nippon Institute of Technology.

Generation of Logical Circuits using Genetics Algorithm and Cellular Automata

This research project's main objective is to automatically synthesize arbitrary logical circuits, using genetics algorithm and cellular automata.

Brief Methodology Explanation:

This research's logical circuits are designed to be analogous to cellular automata.

An array of logical cells are connected to their neighbors, and can perform some logical function. Each cell takes inputs from their neighbors, and outputs to their neighbors. Together, the cell forms some arbitrary logical circuit. We call this device the Logical Cell Array.

Cellular Automata is used to generate a setting for the logical cell array, which is equivalent to using CA to generate random logical circuits. Meanwhile, we use Genetics Algorithm to optimize/search for some desired circuit, and by extension, some desirable Cellular Automata rule.

See the wiki for more info.

---

Brief Usage Explanation:

Programming the FPGA

Compile FPGA project in Quartus Prime.

• Programming via JTAG
  1. Set the DE0-Nano-SoC's MSEL switch to 101010
  2. Using Quartus Programmer, program the FPGA with a .sof file.

Refer to Terasic's DE0-Nano-SoC - User Manual (chapter 3) for a step-by-step guide.

• Programming via EPCS
  1. Set the DE0-Nano-SoC's MSEL switch to 101100
  2. Convert a .sof to a .jic file.
  3. Using Quartus Programmer, program the EPCS device with a .jic file

Refer to Terasic's DE0-Nano-SoC - User Manual (chapter 8) for a step-by-step guide.

Using the GA Program

0. Clear old object files with make clean

1. Cross-compile for ARM Cortex A9 with make arm

Or compile for PC debugging with make pc

2. Copy binary file ga.prog to DE0-nano-SoC with scp

scp <local directory>/ga.prog <Username>@<Remote IP>:/<Remote Directory>

3. Run the binary file in terminal

4. Navigate the program's CLI by entering option numbers.

---

Results Example:

Using Cellular Automata, we generate a 64x64 integer matrix of range [0,3]. This number matrix represents a setting for the logical cell array, and by extension, some arbitrary logical circuit. We can visualize this number matrix, by creating a 64x64 image, where each pixel represents a cell, and each color represents a cell state.

Visualized numerical matrix, which forms a logical OR gate

Here, each color represents the following:

Color	Circuit Equivalence
Black	GND (Cell outputs constant 0)
Cyan	Straight Wire (Cell passes its input 1)
Magenta	Slanted Wire (Cell passes its input 2)
Yellow	NAND Gate (Cell NANDs its input 1 & 2)

The above visualization can be converted to the following circuit image:

Converted logical circuit image, which forms a logical OR gate