Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

DOVA: Deliberation-First Multi-Agent Orchestration for Autonomous Research Automation

Aaron Shen, Alfred Shen

Abstract

Large language model (LLM) agents have demonstrated remarkable capabilities in tool use, reasoning, and code generation, yet single-agent systems exhibit fundamental limitations when confronted with complex research tasks demanding multi-source synthesis, adversarial verification, and personalized delivery. We present DOVA (Deep Orchestrated Versatile Agent), a multi-agent platform introducing three key innovations: (1) deliberation-first orchestration, where explicit meta-reasoning precedes tool invocation, informed by a persistent user model and entity-aware conversation context; (2) hybrid collaborative reasoning, a composable three-phase pipeline unifying ensemble diversity, blackboard transparency, and iterative refinement; and (3) adaptive multi-tiered thinking, a six-level token-budget allocation scheme that reduces inference cost by 40-60% on simple tasks while preserving deep reasoning capacity. We formalize the core algorithms, present an architectural ablation study across seven system configurations, and analyze the contribution of each component to answer confidence, source coverage, and token efficiency.

DOVA: Deliberation-First Multi-Agent Orchestration for Autonomous Research Automation

Abstract

Large language model (LLM) agents have demonstrated remarkable capabilities in tool use, reasoning, and code generation, yet single-agent systems exhibit fundamental limitations when confronted with complex research tasks demanding multi-source synthesis, adversarial verification, and personalized delivery. We present DOVA (Deep Orchestrated Versatile Agent), a multi-agent platform introducing three key innovations: (1) deliberation-first orchestration, where explicit meta-reasoning precedes tool invocation, informed by a persistent user model and entity-aware conversation context; (2) hybrid collaborative reasoning, a composable three-phase pipeline unifying ensemble diversity, blackboard transparency, and iterative refinement; and (3) adaptive multi-tiered thinking, a six-level token-budget allocation scheme that reduces inference cost by 40-60% on simple tasks while preserving deep reasoning capacity. We formalize the core algorithms, present an architectural ablation study across seven system configurations, and analyze the contribution of each component to answer confidence, source coverage, and token efficiency.
Paper Structure (28 sections, 1 theorem, 9 equations, 2 figures, 6 tables, 6 algorithms)

This paper contains 28 sections, 1 theorem, 9 equations, 2 figures, 6 tables, 6 algorithms.

Table of Contents

  1. Introduction
  2. Contributions
  3. Preliminaries
  4. Related Work
  5. System Architecture
  6. Agent Layer
  7. Model Tiering
  8. Core Algorithms
  9. ReAct Reasoning with Self-Reflection
  10. Deliberation-First Orchestration
  11. Hybrid Collaborative Reasoning
  12. Adaptive Multi-Tiered Thinking
  13. Multi-Component Confidence Scoring
  14. Diversity-Aware Memory Retrieval
  15. Query Intent Classification
  16. ...and 13 more sections

Key Result

Proposition 5.1

Let $f_d$ be the fraction of queries where deliberation selects $\texttt{RESPOND\_DIRECTLY}$. The expected tool-call volume relative to a standard ReAct agent is $(1 - f_d)$, achieving cost savings proportional to $f_d \cdot \overline{c}_{\mathrm{tool}}$, where $\overline{c}_{\mathrm{tool}}$ is the

Figures (2)

  • Figure 1: Layered architecture of Dova. Queries enter through the Interface Layer, pass through Orchestration (with deliberation), dispatch to specialized agents, which leverage collaborative reasoning and intelligence services.
  • Figure 2: Token consumption: adaptive vs. fixed Medium. Adaptive saves 94% on classification and 75% on summarization.

Theorems & Definitions (4)

  • Definition 2.1: Agent
  • Definition 2.2: Reasoning Trace
  • Definition 2.3: Confidence Function
  • Proposition 5.1: Tool Call Reduction