Long-form deep research requires multi-faceted investigations over extended horizons to get a comprehensive report. When handling such complex tasks, existing agents manage context at the subtask level to overcome linear context accumulation and information loss. However, they still adhere to a single context window and sequential execution paradigm, which results in mutual interference and blocking behavior, restricting scalability and adaptability. To address this issue, this paper introduces Self-Manager, a parallel agent loop that enables asynchronous and concurrent execution. The main thread can create multiple subthreads, each with its own isolated context, and manage them iteratively through Thread Control Blocks, allowing for more focused and flexible parallel agent execution. To assess its effectiveness, we benchmark Self-Manager on DeepResearch Bench, where it consistently outperforms existing single-agent loop baselines across all metrics. Furthermore, we conduct extensive analytical experiments to demonstrate the necessity of Self-Manager's design choices, as well as its advantages in contextual capacity, efficiency, and generalization.

翻译：长篇深度研究需要在较长时间跨度内进行多方面的调查，以获取全面的报告。在处理此类复杂任务时，现有智能体在子任务层面管理上下文，以克服线性上下文累积和信息丢失问题。然而，它们仍遵循单一上下文窗口和顺序执行范式，这会导致相互干扰和阻塞行为，从而限制可扩展性和适应性。为解决这一问题，本文提出了自管理器，一种支持异步与并发执行的并行智能体循环。主线程可以创建多个子线程，每个子线程拥有独立的隔离上下文，并通过线程控制块对它们进行迭代式管理，从而实现更专注、更灵活的并行智能体执行。为评估其有效性，我们在DeepResearch基准测试中对自管理器进行了评估，结果显示其在所有指标上均持续优于现有的单智能体循环基线方法。此外，我们进行了广泛的分析性实验，以证明自管理器设计选择的必要性，以及其在上下文容量、效率和泛化能力方面的优势。