More on the team

"The challenge lies in solving intricate, real-life problems."

For over 18 years, Avi Harpaz has been working on optimization algorithms that solve large scale problems involving huge numbers of design variables and constraints. For instance, most people step onto an airplane without giving a single thought to the complicated logistic arrangements that go on behind the scenes, before, during and after they are settled into their flight, asking the attendant for a drink. Avi has spent the last few years solving one of the more problematic issues in operational research for the industrial world-the monthly problem of optimizing the pairing of aviation crews with flights.

Optimization problems tackle questions of how to best allocate resources to tasks. These problems come in many shapes and sizes -- from managing workflow in a factory, to assigning parking spaces for planes at an airport, to scheduling pick up and delivery of dairy products. Solving an optimization problem requires three things - generating options to consider, conducting iterative comparisons to decide which options are best given specific criteria, and having the necessary computing processing power to do the iterative comparisons rapidly and on a large scale. The critical sequence of logic used to solve an optimization problem is called an algorithm. Without good algorithms and the computing power to process them, optimization would not be possible.

Another one of Avi's recent projects was optimizing the transportation of dairy products from the manufacturer to the distribution center. The manufacturer needed to know how to best pack the diverse goods into the different sized crates used to transport the dairy products. In addition, the products must be loaded into the trucks so they are easily accessible based on the truck's schedule for pickup and delivery en route. Delivery schedules have to take into account the fact that each truck driver must arrive at a specifically scheduled time to pick up the goods. Also, after the refrigerated truck is loaded, the driver must spend as little time as possible on the road while dropping off dairy goods and picking up empty crates from previous deliveries.

The dairy distribution optimization is a non-standard, extremely complex project that takes into account thousands, even millions of constraints. Avi has yet to deal with a problem that involves a standard textbook solution found in classical optimization models. As a result, he is continuously challenged with developing new algorithms that accurately reflect the many composite problems presented in real-life situations - where there is no option to solve 'everything but...'

When Avi is not busy developing algorithms, he is busy optimizing the amount of quality time he spends with his family, listening to the works of Israeli poets, popular music and watching good movies.