RACE Engine2018-05-25T08:21:50+00:00

RowAnalytics Restraints Engine (RACE)

complex decision support toolkit
  • complex personalized decision support for smart IoT, digital health & dietary advice apps

  • predictably small RAM & CPU footprint on PCs, tablets, smartphones/watches & IoT/sensors

  • delivering complex decision support & personalized advice to users solely on their own devices

  • enabling cheaper, smaller & more secure IoT/wearable solutions with lower power & data needs

  • uses compiled knowledge models describing system behavior via an on-device runtime API

RACE is our patented* Array Logic based Constraint Engine technology. RACE powers our complex data analytics, decision support, smart IoT and embedded sensor/controller solutions.

For many businesses the primary challenge of Big Data is simply too much data. Many such problems can be solved by scaling up the hardware on which analysis tools run using cloud computing. Other solutions require huge amounts of compute power to train deep learning neural nets or to run other machine learning and AI algorithms.

RACE solves the other, fundamentally different problems of modelling complex systems’ behaviors in real-time, where the state of each component in a system may affect the behaviour of others. In these cases, such as personalization of advice in digital health, the problem cannot be simply split across multiple machines (e.g. with Hadoop) as a good solution to one section of the problem may be incompatible with other parts of the problem space.

RACE uses a new branch of maths called Array Logic to unravel the combinatorial complexity of relationships in the system to discover new associations and build high-value predictive models. We build comprehensive models of the problem space, and use the RACE Engine to apply constraints from the real-world to find only those potential solutions that meet all of our criteria. Rather than using machine learning or directed search methods to evaluate all the potential solutions to a problem, we can directly reduce a problem space which might have 101000 potential solutions down to just 1,000, using simple geometrical operators on the logical models. These 1,000 logically valid solutions can then be evaluated and priced for lowest cost/risk and highest benefit.

Critically we can do this in real-time, with a small & predictable memory and CPU footprint and in a provably complete manner (i.e. without taking short-cuts or eliminating any valid solutions) using the RACE API. The RACE API can be deployed across cloud, desktop, mobile, wearable and IoT platforms to power a huge range of applications.

* The application of Array Logic to constraint problems was developed by Dr Gert L. Møller, and our patented RACE technology (US 6,633,863 / EU 1,062,603) applies the mathematical foundation of Array-Based Logic.