Integrating automated body condition scores and lactation data via optimization algorithms to maximize milk revenue...

The objective of this study was to characterize the trade-off between milk revenue and conception costs and to identify BCS profiles during lactation that jointly maximize milk revenue while minimizing conception costs in dairy cows by using the Non-Dominated Sorting Genetic Algorithm II (NSGA-II). A retrospective observational analysis was performed using milk yield, monthly average values from daily BCS automatically recorded after every milking, and fertility records from 12,042 lactations of 7,626 Holstein cows from a commercial dairy farm located in Colorado. Overall, lower BCS was associated with greater milk yield, whereas higher BCS was associated with improved fertility, as indicated by shorter calving to conception intervals, and greater pregnancy probability. In addition, we identified BCS at nadir (the lowest BCS observed during lactation) and monthly BCS values across the first 10 mo of lactation that consistent with efficient outcomes within a lactation, rather than a single optimal trajectory. The NSGA-II analyses revealed that increased milk revenue often comes at the expense of higher costs from delayed conception. Nevertheless, overall profitability is primarily driven by revenue gains from milk production. The monthly BCS average for the most profitable BCS values ranged from 2.50 to 3.27. These scores were relatively lower compared with the least profitable solutions that preserved BCS and had lower delayed conception costs. Consequently, net profit (milk yield revenue - cost of delayed conception) for the most profitable solutions exceeded the least profitable by $1,254, $1,500, $979, and $1,140 per lactation in parity 1 through 4, respectively. The Pareto frontier generated by NSGA-II offers a menu of strategies, from high yield options that accept greater conception costs to balanced approaches that safeguard fertility and herd productivity. These data driven insights advance the use of 2 sourced of highly frequent data points measured at dairy farms (individual milk yield and automated BCS) by integrating economic modeling with biological performance metrics, enabling profit oriented BCS management tailored to lactation stage and economic priorities.