ADVANCED TECHNOLOGIES
Artificial intelligence and optimisation algorithms powering the supply chain and logistics platform.
DATA EXTRACTION USING OCR
Logistimo application uses Optical Character Recognition (OCR) to extract handwritten characters from paper forms. This technology is based on a deep learning algorithm called Convoluted Neural Network (CNN) that can work fully offline on a mobile phone application. The technology enables field workers to efficiently capture data by simply taking a photograph of the paper form. Assistive technology guides one to take good quality photographs. OCR technology extracts handwritten characters. Subsequently, error detection and correction algorithms act on the extracted data to correct errors automatically where possible. Subsequently, the results are presented to the user for a review. Corrective feedback by a community of users is used by the algorithm to learn and create more accurate extraction models.
INVENTORY OPTIMISATION
Logistimo’s SCM platform incorporates an optimisation algorithm that factors forecasted demand, supply and transportation constraints to recommend optimal order quantities. This optimises service levels at minimal cost. Service levels can be specified by inventory item and statistical algorithms are used for forecasting demand and computing optimal quantities at the relevant times. Stock rebalancing and rationing algorithms help in optimising stock as per service levels and consumption in the event of short supply. The algorithm can scale to a large data sets across millions of inventory items.
VEHICLE ROUTING AND SCHEDULING
Vehicle routing and scheduling uses combinatorial optimisation and integer programming algorithms to optimise loads and routes that a fleet of vehicles take. The loads and routes are generated to satisfy delivery SLAs while factoring a variety of constraints including transportation capacity, time/cost of travel, consignment volumes, consignment types, and so on. Routes can be dynamically altered in real-time based on new orders that require pickups en route. The algorithm can scale to across a large number of orders, vehicles, and pickup or delivery points.
