AI Safety

Bio-inspired routing algorithm puts solidarity before distance in São Paulo

This genetic algorithm cut route length variation by 89% while prioritizing worker equity.

Deep Dive

A new research paper from Gustavo N. Gonçalves, Cristiano C. Cruz, and Romis Attux presents a routing algorithm built for Señoritas Courier, a bicycle delivery cooperative in São Paulo that employs exclusively cis women and trans people. Unlike traditional logistics optimization that minimizes distance or cost, this cooperative operates on principles of solidarity, care, and equitable income distribution. The researchers found that the classical Vehicle Routing Problem (VRP) was inadequate because it ignores individual constraints and fairness. They formulated a new variant called the Señoritas Routing Problem, which incorporates biker-specific limits on weight, volume, and maximum distance, alongside a solidarity objective that balances route lengths.

To solve this problem, the team employed a genetic algorithm and compared three fitness formulations: a baseline distance-minimization operator, a constrained version, and a progressive formulation that penalizes workload imbalance. The progressive constrained formulation eliminated all constraint violations and slashed the standard deviation of route lengths from 7.92 km to 0.81 km—a 90% reduction—at the cost of a moderate increase in total distance. The algorithm was developed through a participatory process where cooperative members acted as co-designers, ensuring the technical solution aligned with their values and needs. This case demonstrates that operations research can be reoriented toward solidarity, and that participatory methodologies are essential for worker cooperatives.

Key Points
  • Señoritas Courier is a São Paulo bicycle delivery cooperative run exclusively by cis women and trans people, prioritizing solidarity over pure efficiency.
  • The progressive constrained genetic algorithm reduced route length standard deviation from 7.92 km to 0.81 km while eliminating all constraint violations.
  • The algorithm was co-designed with cooperative members, ensuring the technical solution reflects their values of fairness and equitable income distribution.

Why It Matters

Shows AI can optimize for equity and solidarity, not just efficiency—offering a blueprint for ethical logistics.

📬 Get the top 10 AI stories daily