An agent-based simulation model to compare different reproductive strategies in cow-calf operations: Technical performance

The objective of this study was to create a stochastic, agent-based simulation model of a synthetic population of beef cattle, and then use it to compare the technical performance of different reproductive strategies. The model was parameterized using data from a real beef cattle herd and from the peer-reviewed scientific literature to represent a Nelore cattle herd in the state of Sao Paulo, Brazil. Ten scenarios were evaluated: natural mating (NM) only (ONM); one timed artificial insemination (TAI) plus NM (1TAI + NM); two TAI plus NM, with 24, 32, and 40 days between inseminations (2TAI/24 + NM, 2TAI/32 + NM, and 2TAI/40 + NM, respectively); three TAI without NM, with 24, 32, and 40 days between TAI (3TAI/24, 3TAI/32, and 3TAI/40, respectively); and three TAI plus NM, with 24 and 32 days (3TAI/24 + NM and 3TAI/32 + NM, respectively). NM began 10 days after the last TAI and was performed until the end of the breeding season. The size of the female herd was set to contain up to 400 individuals. The bull population was established at 0, 7, or 15 bulls depending on the used scenario. Simulation was performed for 5000 days. The outcomes for each scenario are means ± S.E. assessed on 32 farms at 1-day time intervals and on an animal-by-animal basis after steady state was reached (1825 days). The 3TAI/24 + NM scenario resulted in a greater number of births (279.85 ± 0.47 births), while the ONM scenario had the least value (202.38 ± 0.43 births). The heaviest males and females at weaning belonged to 3TAI/24, with 190.85 ± 0.17 kg for males and 173.89 ± 0.13 kg for females. The ONM scenario had the lightest males (166.84 ± 0.18 kg) and females (151.75 ± 0.16 kg). The greatest and least total pregnancy rates were found in 3TAI/24 + NM (0.91 ± 0.00) and ONM (0.62 ± 0.00), respectively. The ONM scenario required 52.5 days more than scenarios that included TAI to reach 50% of pregnancy. The greatest ages at culling for cows was 3TAI/24 + NM (3658.88 ± 10.41 days). In contrast, the lowest age at culling was found in ONM (2823.93 ± 8.28 days). We concluded that the proposed model represents the main interactions of a real beef cattle herd. It has all the advantages of a physical experiment, but does not require incurring significant expenses nor altering the real system. This study offers evidence that the scenarios that present the best technical performance are those that used TAI with a 24-day interval between inseminations.