The p-median solver developed by the Track&Know project uses a genetic algorithm approach to solve the Location-allocation problems in polynomial time. Location-allocation problems typical deal with provisioning of resources between facilities based on historic demand. The p-median approach is one such model that aims to minimize the total demand-weighted distance between the demand points and the facilities. This NP-Hard problem aims to locate ‘p’ facilities to serve ‘n’ demand, by minimizing the total demand-weighted distance between the facilities and the demand.
This tool takes as inputs coordinate pairs and makes location-allocation allocation decisions based on geodesic, road distance, or journey time constraints.
- Containerized solution for seamless rapid deployment
- Apache Spark support for big data applications
- Bundled with a GUI for ease of use
- AI driven with genetic algorithms
- Interfaces for open route service and others for true travel time and distance calculations
Location-allocation problems can be applied to a multitude of areas where finite resources and services have to be delivered to a geographically disparate clientele. From manufacturing, warehousing to healthcare service positioning, the p-median method can help businesses and service providers locate their facilities such that they can maximize access and access times for their customers.
In a post COVID-19 world, with the increase of e-commerce for basic household items, re-distribution of warehouses to optimize ‘next day’ or even ‘next hour’ deliveries is a game changer for national and regional stores. Electric mobility and electric last mile delivery are all enabled by optimally locating facilities in accordance with demand points by minimizing the travel distance. This tool can also play an important role in translating mobility information of citizens into policy and management recommendations e.g. placement of vehicle charging facilities, or out-reach / popup healthcare services.
This tool unlike others on the market is free and opensource allowing end users to assess the methods employed in making location allocation decisions. Decisions can therefore be taken with confidence that any bias has been minimized and no sector of the population is being disenfranchised.
Additionally, the tool is simple to use with an intuitive GUI and can work with just a list of coordinates. For larger datasets the Apache spark variant can be used to leverage big data and processing infrastructure.
- Uses Genetic Algorithms
- Containerised for web, desktop, or server use
- Includes sample data for verification