Data and Code for: Enhanced Memetic Algorithm with Adaptive Local Search for Area-Optimized Standard-Cell Placement in VLSI Physical Design
Description
This dataset accompanies the research article “Enhanced Memetic Algorithm with Adaptive Local Search for Area-Optimized Standard-Cell Placement in VLSI Physical Design”. The dataset contains: 1. Synthetic benchmark instances — Ten synthetic placement benchmark configurations (m = 3–49 modules), including module dimensions, netlist connectivity, and slot assignments used in Table 1 of the manuscript. 2. MCNC/GSRC benchmark circuits — Input files for the industrial benchmark circuits apte, xerox, and hp used in Table 2, obtained from the public ICCAD placement benchmark suite. 3. Experimental result tables — Raw numerical results from 30 independent runs for all benchmarks, including bounding-box area, fitness values, wirelength, convergence generation counts, and runtime measurements corresponding to Tables 1–4. 4. EMA source code — ISO C implementation of the Enhanced Memetic Algorithm (EMA), compiled using GNU GCC under Linux with -O2 optimization, including adaptive local search, elitist replacement, and chromosome encoding modules. 5. README documentation — Instructions for compilation and execution, benchmark file formats, parameter settings, and reproduction guidelines.
Files
Steps to reproduce
1. Download and unzip ema_source_code.zip. Compile with: gcc -O2 -o ema ema.c -lm 2. Unzip benchmarks.zip to obtain the 10 synthetic (.txt) and 3 MCNC circuit input files. 3. Run EMA on a benchmark: ./ema benchmarks/syn_49.txt --runs 30 --seed 42 4. For 30-run statistical evaluation (reproducing Tables 1–2): ./ema benchmarks/syn_49.txt --runs 30 5. Parameters fixed across all experiments: P=20, Gmax=200, pc=0.85, pm=0.10, α=0.10, LSmin=5, LSmax=10. 6. Results are written to stdout and match the raw data in results_tables.xlsx. 7. See README.txt for full compilation instructions, file format descriptions, and parameter guidance. Environment: Linux, GNU GCC, Intel Core i5 3.0 GHz, 8 GB RAM (standard desktop workstation).