Intelligence-Driven Optimization of Smart Contracts

In this chapter, we present an overview on how smart contracts could be optimized by intelligence-driven approaches. We empirically study the repetitiveness of smart contracts via cluster analysis and try to extract differentiated codes from the similar contracts. Differentiated codes are defined as the source codes except the repeated ones in two similar smart contracts, usually illustrating how a software feature is implemented or a programming issue is solved. Then, a differentiated code might be used to guide the update of a smart contract in its next version. Finally, we apply syntax and semantic similarities to discover its similar smart contracts from more than 120,000 smart contracts, and recommend the differentiated codes to the target smart contracts.