An Automated Multi-Agent Search-then-Evolution Framework for Lifetime Value Prediction

Abstract

Lifetime value (LTV) prediction is a key component of decision-making in online advertising, e-commerce, and recommender systems. In practice, however, LTV modeling is scenario-dependent because data distributions differ across applications. Practitioners often rely on scenario-specific pipelines with repeated feature engineering, objective adjustment, and tuning. These workflows require manual effort and are hard to transfer. We propose AgentLTV, an automated multi-agent search-then-evolution framework for lifetime value prediction. AgentLTV treats each candidate solution as an executable pipeline program. LLM-driven agents generate code, run and repair pipelines, and analyze execution feedback. Two decision agents coordinate a two-stage search. The Monte Carlo Tree Search (MCTS) stage explores a broad space of modeling choices under a fixed budget, guided by the Polynomial Upper Confidence bounds for Trees criterion and a Pareto-aware multi-metric value function. The Evolutionary Algorithm (EA) stage refines the best MCTS program via island-based evolution with crossover, mutation, and migration. Experiments on a large-scale proprietary dataset and a public benchmark show that AgentLTV consistently discovers strong models across ranking and error metrics. Online bucket-level analysis further indicates improved ranking consistency and value calibration, especially for high-value and negative-LTV segments. We summarize practitioner-oriented takeaways: use MCTS for rapid adaptation to new data patterns, use EA for stable refinement, and validate deployment readiness with bucket-level ranking and calibration diagnostics. The proposed AgentLTV has been successfully deployed online. Details of our optimal model and prompts are available at https://github.com/wuchaowei123/AgentLTV.