A Decoupled Basis-Vector-Driven Generative Framework for Dynamic Multi-Objective Optimization

2026-04-01

AI Summary

This work addresses the challenges in dynamic multi-objective optimization where nonlinear coupling, negative transfer from historical data, and cold-start effects during environmental changes hinder effective tracking of the Pareto front. To overcome these issues, the paper proposes the DB-GEN framework, which uniquely integrates decoupled basis vectors with generative modeling. By employing discrete wavelet transform to decouple evolutionary trajectories and sparse dictionary learning to construct transferable basis vectors, DB-GEN generates a structured latent manifold under topological-aware contrastive constraints. A surrogate model further enables zero-shot, rapid initialization without fine-tuning. The method achieves online adaptation within milliseconds (approximately 0.2 seconds) after environmental shifts and significantly improves Pareto front tracking accuracy across multiple dynamic benchmarks, effectively mitigating both negative transfer and cold-start problems.

Abstract

Dynamic multi-objective optimization requires continuous tracking of moving Pareto fronts. Existing methods struggle with irregular mutations and data sparsity, primarily facing three challenges: the non-linear coupling of dynamic modes, negative transfer from outdated historical data, and the cold-start problem during environmental switches. To address these issues, this paper proposes a decoupled basis-vector-driven generative framework (DB-GEN). First, to resolve non-linear coupling, the framework employs the discrete wavelet transform to separate evolutionary trajectories into low-frequency trends and high-frequency details. Second, to mitigate negative transfer, it learns transferable basis vectors via sparse dictionary learning rather than directly memorizing historical instances. Recomposing these bases under a topology-aware contrastive constraint constructs a structured latent manifold. Finally, to overcome the cold-start problem, a surrogate-assisted search paradigm samples initial populations from this manifold. Pre-trained on 120 million solutions, DB-GEN performs direct online inference without retraining or fine-tuning. This zero-shot generation process executes in milliseconds, requiring approximately 0.2 seconds per environmental change. Experimental results demonstrate that DB-GEN improves tracking accuracy across various dynamic benchmarks compared to existing algorithms.

Problem

• dynamic multi-objective optimization
• non-linear coupling
• negative transfer
• cold-start problem
• Pareto front tracking

Innovation

• decoupled basis-vector
• dynamic multi-objective optimization
• sparse dictionary learning
• latent manifold
• zero-shot generation