Article: Adaptive genetic algorithm for multi-topic polyphonic music generation Journal: International Journal of Information and Communication Technology (IJICT) 2025 Vol.26 No.45 pp.83 - 100 Abstract: Automatic composition struggles to balance multi-theme planning with strict polyphonic constraints. To address this, this paper proposes an adaptive genetic framework for multi-topic polyphonic music generation. In the scheme, first a bar-aligned, voice-aware representation is prepared with tonal cues and a theme schedule. Then a domain-aware evolutionary core explores the search space through voice-preserving crossover, musically constrained mutation, and lightweight local repair around exposed phrases. Finally, a composite evaluator guides selection while an adaptive controller adjusts operator rates using diversity and stagnation signals. Experiments on chorale, chamber, and modern tonal sets show fewer rule violations, higher consonance with 83% vertical consonance and tonal stability, stronger theme recognisability, and faster convergence without extra runtime. The approach delivers structured, stylistically credible music with strong controllability, clear diagnostics, and room for interactive use. Inderscience Publishers - linking academia, business and industry through research

Title: Adaptive genetic algorithm for multi-topic polyphonic music generation

Authors: Huiyan Zeng

Addresses: School of Tourism Home Economics and Arts, Qingyuan Polytechnic, Qingyuan, 511500, China

Abstract: Automatic composition struggles to balance multi-theme planning with strict polyphonic constraints. To address this, this paper proposes an adaptive genetic framework for multi-topic polyphonic music generation. In the scheme, first a bar-aligned, voice-aware representation is prepared with tonal cues and a theme schedule. Then a domain-aware evolutionary core explores the search space through voice-preserving crossover, musically constrained mutation, and lightweight local repair around exposed phrases. Finally, a composite evaluator guides selection while an adaptive controller adjusts operator rates using diversity and stagnation signals. Experiments on chorale, chamber, and modern tonal sets show fewer rule violations, higher consonance with 83% vertical consonance and tonal stability, stronger theme recognisability, and faster convergence without extra runtime. The approach delivers structured, stylistically credible music with strong controllability, clear diagnostics, and room for interactive use.

Keywords: algorithmic composition; polyphonic music; adaptive genetic algorithm; thematic scheduling.

DOI: 10.1504/IJICT.2025.150410

International Journal of Information and Communication Technology, 2025 Vol.26 No.45, pp.83 - 100

Received: 30 Aug 2025
Accepted: 02 Oct 2025
Published online: 12 Dec 2025 *

Title: Adaptive genetic algorithm for multi-topic polyphonic music generation

Keep up-to-date