#12 Research meet-up: Using optimisation to improve market access for Brazilian smallholder farmers

Andrea Tuni presented a decision-support system designed to support smallholder farmers in deciding which institutional-market contracts to pursue and how to distribute their produce. The approach was tested in the Canudos settlement in Goiás and is being transferred into web-based applications for wider practical use.

Andrea Tuni is an Assistant Professor in the Department of Management and Production Engineering at Politecnico di Torino in Italy. His research covers sustainable supply chains, circular economy, reverse logistics, performance measurement, food distribution and applications of solidarity-economy models.

Why can access to institutional markets be difficult?

Smallholder farmers can face unstable incomes when access to reliable markets is limited.

In Brazil, institutional markets were created to allow smallholder farmers to supply produce to schools through a non-competitive bidding mechanism. Instead of competing through a conventional auction, farmers can respond to public purchasing opportunities under defined conditions.

Participation can still be difficult because farmers need to determine whether a particular call will be profitable and then coordinate the distribution of produce if they receive a contract.

Why is this a prescriptive analytics problem?

The decision-support system addresses several connected choices: which opportunities to bid for, which contracts to accept and how deliveries should be organised to maximise farmers' returns.

For distribution planning, it uses a Heterogeneous Fleet Split Delivery Vehicle Routing Problem. In practical terms, this is an optimisation model that determines how different types of vehicles should make deliveries and allows an order to be divided between more than one vehicle or trip when useful.

How is the approach being tested?

The system was applied to the Canudos settlement in Goiás state. The project is also progressively embedding the decision-support functions in Progressive Web Apps — websites designed to work more like mobile applications — so that the analytical methods can be used more easily outside a research environment.