Shri Shakti LPG used integer programming to place bottling plants

Shri Shakti LPG imported and marketed propane in South India, selling it in cylinders to domestic and commercial customers through dealers. Dealers refilled from bottling plants, and those plants drew bulk LPG from the cheaper of two coastal import-and-storage facilities.

The tempting approach was to choose plant sites and sizes from managerial intuition, local demand and transport convenience. Sankaran and Raghavan instead implemented integer programming to decide the locations and long-run sizes of the bottling plants together.

The Interfaces abstract estimates that the recommended bottling-plant configuration was about $1 million cheaper annually than the configuration Shri Shakti LPG had initially planned.

Size the plants with the map

The model made import terminals, bottling capacity and dealer demand part of the same cost equation.

Choosing size and location together avoided building cheap plants in expensive network positions.

Integer programming let managers test a capital plan before concrete was poured.

Facility location is not a real-estate question alone; plant size, inbound supply and dealer replenishment have to be optimized as one system.

The case was published in Interfaces as a petroleum and natural-gas facility-location application, with RePEc indexing it under discrete location and equipment-planning keywords.