Achieving Privacy-Preserving Optimizer Architecture for Intent Execution on EVM Blockchain

Intention adoption represents a migration from an imperative to a declarative paradigm, which is expected to significantly improve the user experience in blockchain. While the development of account abstraction has expanded the range of possibilities for intent expression, the architecture for intent-centric collaboration and coordination remains underdeveloped. Existing works depend on trusted centralized entities and employ broadcasts to the network to disseminate user intent, risking single points of failure and privacy leakage. In addition, no standard intent architecture has been proposed. In this paper, we propose a novel decentralized privacy-preserving optimizer architecture for executing intents on the EVM blockchain. We first propose standard intent structures and zero-knowledge intent structures. Our architecture allows users to transform raw intent structures into zero-knowledge intent structures and send them to a set of trusted optimizers. The optimizers can optimize and execute the intents without revealing the content of the intents. We use zero-knowledge proof and public-key encryption to ensure the privacy and correctness of the intents, and smart contracts to ensure the security and verifiability of the execution. We also design a quantization function to evaluate and compare solutions from different optimizers. We analyze the security and performance of our architecture and show that it can achieve privacy, efficiency, and scalability of intent execution on the EVM blockchain.