Flexible and Trustworthy Execution of Business Processes

Blockchain technology has emerged as a transformative solution for execut- ing inter-organizational business processes in distributed environments where trust between participants cannot be assumed. By encoding business logic into smart contracts, blockchain ensures tamper-proof and transparent exe- cution according to predefined rules. The Business Process Model and No- tation (BPMN) standard, particularly its choreography diagrams, provides a high-level, intuitive language for modeling these multi-party interactions. However, a significant challenge arises from the inherent immutability of blockchain, which conflicts with the need for business processes to adapt to unforeseen circumstances at runtime. Traditional platforms like Ethereum require complex, error-prone, and costly patterns (e.g., proxy contracts) to enable updates, undermining both flexibility and trust. This thesis addresses this challenge by leveraging the Algorand blockchain, which natively supports the secure and efficient updating of stateful smart contracts. We propose a novel, end-to-end framework for the flexible and trustworthy execution of BPMN choreographies. The core con- tributions are fivefold. First, we present a formal, modular, and semantics- driven method for translating BPMN choreography models into smart con- tract code, ensuring a one-to-one correspondence between model elements and contract functions. Second, we detail a mechanism for runtime adap- tation, where changes to the choreography model are seamlessly reflected in the underlying Algorand smart contract using its native update capabil- ities, eliminating the need for complex patterns. Third, we explore the use of Large Language Models to assist in migrating existing Solidity contracts to Algorand, reducing the barrier to entry. Fourth, we introduce a BPMN to Algorand smart contract generation tool that automates the translation of BPMN choreography models into executable smart contracts, simplifying x the development process and eliminating the need to write smart contracts from scratch. Finally, by exploiting our BPMN-to-Algorand translation tool we validate our approach through experiments and case studies, demonstrat- ing its feasibility, cost-effectiveness, and superior performance compared to Ethereum-based solutions. The results confirm that the proposed Algorand- based approach provides a robust foundation for achieving both trustwor- thiness through decentralization and immutability, and flexibility through secure and straightforward contract updates, thereby effectively supporting dynamic business processes.