BALANCING THE WORKLOAD OF THE TRAVELING SALESMAN OF COWS ARTIFICIAL INSEMINATION SERVICES

In Israel about 1090 inseminations are performed each day, in 412 farms at 283 sites. This paper presents tools, developed to plan agent work schedules which are implemented in the artificial insemination services in Israel. The tools developed are: (1) a multi-variable regression model to predict the duration of cows’ treatment; (2) an optimization model of farm allocation and routes determination, and (3) solution schemes for this model. The solution consists of four steps: (1) heuristic algorithms for initial allocation of customers to agents; (2) improvement of the total traveling time by neighbor search; (3) improving average daily working time by neighbor search and (4) applying a genetic algorithm to determine the desired route for each worker. The result of work planning and an examination of the trade-off between work balancing and the total traveling time are presented, too.