Optimization objectives

Every time an order is sent to IBM® Sterling Intelligent Promising, it performs multi-objective optimization to determine the best option for fulfilling it.

Depending on the type of offering, Sterling Intelligent Promising uses two different optimization approaches to consider the business optimization objectives when optimizing orders to lower fulfillment costs and use inventory more efficiently:
Deterministic optimization
This is a rules-based optimization approach that calculates the lowest cost-to-serve using static variables. It is available with the Sterling Intelligent Promising Standard offering. The following features are included:
  • Minimizing shipping cost
  • Balancing node capacity
  • Minimizing processing cost
  • Optimizing delivery speed
Predictive optimization
This approach uses Machine Learning (ML) models to evaluate probabilistic variables. It is available with the Sterling Intelligent Promising Premium offering. The following features are included:
  • Avoiding markdowns
  • Avoiding stockouts
  • Optimizing node performance (Preview only)

Each optimization objective is configured as a weight that ranges from 0% to 100% to determine the priority in cost optimization. IBM Sterling Intelligent Promising prioritizes objectives with higher weights. Setting all objectives to the same value or to 100% results in balanced optimization across all levers. To disable an objective, set the weight to 0%.

Note: In the Display settings page, select the General tab and browse to the Promise and fulfill section to toggle the Optimizing node performance setting. Use this setting to show or to hide the Optimizing node performance setting in the Optimization objective page. Also, use this setting to show or to hide the Optimizing delivery speed setting in the Optimization dashboard and the Optimization explainer Order summary page.

You set weights for each of these objectives based on your business goals. Then, Sterling Intelligent Promising incorporates those weights into the following equation to determine the best option for fulfillment based on your business priorities:
(Shipping cost × Optimization weight) + (Capacity penalty × Optimization weight) + (Processing cost × Optimization weight)+ (Stockout avoidance cost × Optimization weight) - (Markdown avoidance cost × Optimization weight) + (Node performance cost × Optimization weight) + (Transit time cost × Optimization weight)