EN
Back to the archive

The encyclopedia · Software & IT · Technical decision · 1985–1987

Mobil's one-second dispatch optimizer ran $4B of fuel and saved $3M a year

Mobil's CAD used integer programming — about one second per dispatch — across 600,000 orders a year; dispatcher productivity more than doubled.

Mobil Oil Corporation

the move

In spring 1985 Mobil began full operation of a nationwide system for dispatching and processing orders for gasoline, diesel and heating oil. The operation was huge: $4 billion in annual sales, 600,000 customer orders, 120 bulk terminals and a fleet of more than 430 vehicles.

The heart of the system was computer-assisted dispatch (CAD), a collection of integer-programming methods embedded in a real-time, transaction-driven information system. It allots about one second, on average, per dispatch for optimization, exploiting the experience of the human dispatchers it assists while uniformly enforcing management policy.

Results were measured and reported by Mobil itself: dispatcher productivity more than doubled on orders delivered per person; annual net cost savings for product distribution exceeded $2 million; and operating savings for light products reached about $3 million a year, with dispatching centralized at a single Light Products Control Center.

why it works

  • A one-second solve let optimization run every dispatch, not just big ones.
  • Human dispatcher knowledge was encoded as the model's starting point.
  • Central control enforced pricing and delivery policy uniformly.
  • Scale — 600,000 orders a year — made seconds-per-dispatch pay.
the payoffOptimize each dispatch in about one secondinspired

what transfers

Automate the routine decision so fast it disappears: a one-second solve keeps operators in control while the optimizer enforces policy — productivity doubled, savings hit millions a year.

what came after

Mobil's CAD became a landmark of real-time OR practice, and the paper by the company's team with Naval Postgraduate School researchers is still cited as an early model of optimization embedded in operations. The centralization it enabled reduced costs and staff while improving customer service.

references

spotted an error? The archive wants to know.

same kind of clever