In pig production, the reproductive phase represents a critical point for farm profitability. Currently, the assignment of semen doses to sows ready for insemination is performed randomly or based exclusively on the experience of farm staff. This practice does not systematically consider the multiple variables that influence reproductive success, such as semen characteristics, the reproductive history of the sow, or the optimal timing of the reproductive cycle. This randomness in decision-making introduces great variability, which can directly affect the efficiency and economic sustainability of farms.
At the Universidad de Murcia (UMU), a predictive model based on machine learning techniques has been developed that analyzes real data from the pig reproductive process.
Its objective is to identify the male–female combinations with the highest probability of success based on variables such as sperm quality and ejaculate preservation, the physical and reproductive status of the sow, and the conditions of the insemination process; thus optimizing pairings at the time of artificial insemination.
The practical application of this model makes it possible to transform a traditionally random process into a data-based strategic decision, improving productivity (more piglets per litter, higher average litter weight), reducing variability and increasing farm profitability. The system generates a ranking of optimal pairings between the available semen doses and the sows ready for insemination.
Benefits:
- Increase in reproductive performance through the optimal assignment of boars and sows.
- Improvement in semen use efficiency by avoiding pairings with a low probability of success.
- Reduction in variability in production results thanks to the use of historical data and predictive models.
- Support for on-farm decision-making through objective rankings of reproductive pairings.
The objective of the contact is to obtain commercial feedback on the technology to guide it towards market needs and seek a collaboration that leads to the exploitation of the proposed system.
Institution: Universidad de Murcia
TRL: The model is in the development and validation phase. The databases are being expanded and the models adjusted to improve their accuracy and applicability.
Funding: Prueba de Concepto – Fundación Séneca
Contact: Ana Carlota de la Cruz Abad / a.cruz@viromii.com
Interested in this opportunity?
Contact us to discuss licensing, collaboration or the next steps.
Contact project