MORetro*: Multi-Objective Algorithm for Pareto-Optimal Synthesis Planning

Researchers have introduced MORetro*, a multi-objective search algorithm for computer-aided synthesis planning that generates a Pareto front of synthesis routes. Unlike traditional methods that focus on a single feasible route, MORetro* explicitly balances competing objectives like cost, sustainability, toxicity, and yield. The algorithm uses weighted scalarization and Bayesian optimization-informed sampling to navigate combinatorial search spaces efficiently. Building on multi-objective A*-search, MORetro* provides optimality guarantees, recovering the true Pareto front for a fixed single-step model. The work was published on arXiv (2605.07521) and addresses a misalignment between current CASP methods and real-world chemical practice.