Sequential Multi-Agent Dynamic Algorithm Configuration

2025-10-27

AI Summary

Existing automated configuration methods overlook inherent inter-parameter dependencies in complex algorithms—e.g., operator type constraining admissible parameter values. To address this, we propose Seq-MADAC, a sequential multi-agent dynamic configuration framework. Its core innovation is the first introduction of a dependency-aware parameter action ordering mechanism, instantiated via a Sequence Advantage Decomposition Network (SADN) that explicitly models the execution order of parameter adjustments. Seq-MADAC integrates multi-agent reinforcement learning with order-sensitive policy optimization to enable dynamic hyperparameter tuning. Evaluated on synthetic function optimization and multi-objective algorithm configuration tasks, Seq-MADAC significantly outperforms state-of-the-art methods in both convergence speed and solution quality. Moreover, it demonstrates strong generalization across diverse problem classes, validating its robustness to structural variations in algorithmic configurations.

Technology Category

• Search and Optimization: Algorithm Configuration
• Multiagent Systems: Adversarial Agents
• Reasoning under Uncertainty: Sequential Decision Making

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Dynamic algorithm configuration (DAC) is a recent trend in automated machine learning, which can dynamically adjust the algorithm's configuration during the execution process and relieve users from tedious trial-and-error tuning tasks. Recently, multi-agent reinforcement learning (MARL) approaches have improved the configuration of multiple heterogeneous hyperparameters, making various parameter configurations for complex algorithms possible. However, many complex algorithms have inherent inter-dependencies among multiple parameters (e.g., determining the operator type first and then the operator's parameter), which are, however, not considered in previous approaches, thus leading to sub-optimal results. In this paper, we propose the sequential multi-agent DAC (Seq-MADAC) framework to address this issue by considering the inherent inter-dependencies of multiple parameters. Specifically, we propose a sequential advantage decomposition network, which can leverage action-order information through sequential advantage decomposition. Experiments from synthetic functions to the configuration of multi-objective optimization algorithms demonstrate Seq-MADAC's superior performance over state-of-the-art MARL methods and show strong generalization across problem classes. Seq-MADAC establishes a new paradigm for the widespread dependency-aware automated algorithm configuration. Our code is available at https://github.com/lamda-bbo/seq-madac.

Problem

• Addressing inherent parameter interdependencies in dynamic algorithm configuration
• Improving automated tuning of complex algorithms with sequential dependencies
• Enhancing multi-agent reinforcement learning for dependency-aware hyperparameter optimization

Innovation

• Sequential multi-agent framework for dependency-aware algorithm configuration
• Sequential advantage decomposition network leveraging action-order information
• Superior performance across synthetic functions and optimization algorithms