Parameters affecting optimal herd life of Holestein cows in North West of Iran using stochastic dynamic programming

The objective of this study was to determine the optimal herd life of dairy cows in North West of Iran based on data collected from nine large herds of three provinces (Ardabil, East Azerbaijan and West Azerbaijan) and to assess its affecting factors using data on costs and revenues in 2012 and by stochastic dynamic programming. Variables used to describe the state of cows included lactation number at 10 levels, production ability at 3 levels and reproductive performance at 4 levels. Markov chain was used to incorporate the probability for each level of factors in determining the status of a cow. The objective function was the expected net present value of a cow in a 10-years horizon and the optimization criterion was to maximize this function. Using dynamic programming the optimal combination of the state variables was realized and hence the optimum herd life of the cows was estimated to be 4.99 years. Examination of optimal objective function in different states revealed that increasing the production level and reducing the calving interval resulted in higher expected net present value of cows. The effect of changes in price of heifers, milk and discount rates on optimal herd life was examined and it was found that by increasing heifer price, lowering milk price and increasing the discount rate, the culling rate was decreased and optimal herd life was increased.