Method and system for optimizing retail promotions with mixed-integer non-linear programming and cross-elasticity effects

A system for promotion planning, the system comprising:

• a memory storing instructions;
• one or more Input/Output (I/O) interfaces;
• and
• one or more hardware processors coupled to the memory via the one or more I/O interfaces, wherein the one or more hardware processors are configured by the instructions to:
• create a plurality of bundles within a plurality of items for promotion plan at a store group level based on inputs derived from affinity analysis for identifying co-purchased items, a complement addition for augmenting existing bundles, clustering items based on product attributes, and propensity scores, wherein items and bundles are created for each customer segment among a plurality of preidentified customer segments in the store or the store group;
• select by executing a greedy selection technique and a local search heuristic a plurality of optimization candidates comprising a set of items from among the plurality of items and a set of bundles from among the plurality of bundles, and for a promotion optimization problem, wherein the greedy selection technique and the local search heuristics is applied iteratively on the plurality of items and the plurality of bundles arranged in descending order of a greedy score metric and filtered at an initial level based on business rules and bundle conflicts, a total budget of the promotion planning, and the propensity scores to obtain the plurality of optimization candidates;
• and
• identify a promotion list comprising optimal items and bundles and associated optimal discounts from among the plurality of optimization candidates using a Mixed Integer Non-Linear Programming (MINLP) that maximizes anon-linear objective function while satisfying a plurality of constraints, the plurality of constraints comprising a total margin loss constraint, preset minimum and maximum price discount bounds for the items and the bundles, the business rules, the bundle conflicts, cross-price elasticity within a demand function of the item during the MINLP optimization to capture effect of promotions on one item or bundle affecting a demand for other items or bundles, wherein the MINLP maximizes the weighted combination of a total margin and a total revenue while allowing to exceed preset revenue and margin targets and applying constraints for a promotion period.

One or more non-transitory machine-readable information storage mediums comprising one or more instructions which when executed by one or more hardware processors cause:

• creating a plurality of bundles within a plurality of items for promotion plan at a store group level based on inputs derived from affinity analysis for identifying co-purchased items, a complement addition for augmenting existing bundles, clustering items based on product attributes, and propensity scores, wherein items and bundles are created for each customer segment among a plurality of preidentified customer segments in the store or the store group;
• selecting by executing a greedy selection technique and a local search heuristic a plurality of optimization candidates comprising a set of items from among the plurality of items and a set of bundles from among the plurality of bundles, and for a promotion optimization problem, wherein the greedy selection technique and the local search heuristics is applied iteratively on the plurality of items and the plurality of bundles arranged in descending order of a greedy score metric and filtered at an initial level based on business rules and bundle conflicts, a total budget of the promotion planning, and the propensity scores to obtain the plurality of optimization candidates;
• and
• identifying a promotion list comprising optimal items and bundles and associated optimal discounts from among the plurality of optimization candidates using a Mixed Integer Non-Linear Programming (MINLP) that maximizes anon-linear objective function while satisfying a plurality of constraints, the plurality of constraints comprising a total margin loss constraint, preset minimum and maximum price discount bounds for the items and the bundles, the business rules, the bundle conflicts, cross-price elasticity within a demand function of the item during the MINLP optimization to capture effect of promotions on one item or bundle affecting a demand for other items or bundles, wherein the MINLP maximizes the weighted combination of a total margin and a total revenue while allowing to exceed preset revenue and margin targets and applying constraints for a promotion period.