Indian Wedding Optimization algorithm beats GA, PSO on benchmarks

Researchers led by Deepika Saxena (including Kishu Gupta, Jitendra Kumar, Jatinder Kumar, Sakshi Patni, Vinaytosh Mishra, Niharika Singh, Ashutosh Kumar Singh) have introduced Indian Wedding System Optimization (IWSO) — a novel population-based metaheuristic modeled on the socio-cultural dynamics of Indian weddings. The algorithm treats the matchmaking process as a guided, selective search: families, candidates, and matchmakers collaborate to form a framework that solves complex optimization problems. IWSO's first innovation is a matchmaker-guided influence strategy, where elite solutions actively steer weaker candidates, accelerating convergence without requiring external control parameters. The second is an adaptive elimination and reinitialization mechanism that periodically replaces underperforming individuals, preventing premature convergence and preserving population diversity.

Extensive experiments on benchmark high-dimensional and multimodal test functions (e.g., Rastrigin, Ackley, Griewank) demonstrate that IWSO consistently outperforms classical metaheuristics including Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Differential Evolution (DE), and Cuckoo Search (CS) in terms of convergence speed, final solution quality, and robustness across multiple runs. The paper also provides analytically derived time and space complexity, establishing theoretical rigor. While still in preprint on arXiv (dated May 5, 2026), the work suggests that culturally inspired algorithms can match or exceed established optimization techniques, opening the door for more socially grounded heuristics in machine learning and engineering design.