Smart Order Router with ML Prediction Engine

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In electronic order routing, the rapid pace of market updates and the fleeting nature of quotes pose a challenge, as existing systems struggle to process execution instructions quickly enough to avoid stale quotes, leading to undesirable prices or unfilled orders.

Innovation Solution

The implementation of a machine learning approach that dynamically modifies order configurations through heuristic stepwise optimization, utilizing data structures like hash maps to efficiently process and transmit execution instructions within tight time constraints, and incorporating machine learning models to estimate probabilities and risk-reward profiles for optimal order placement and routing.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If traditional order routing systems are used to process execution instructions, then the system can handle basic order routing functions, but the processing speed is insufficient to beat market updates and avoid stale quotes

Engineering Contradiction:

Improveprocessing speed of execution instructionsVSAvoidtime window before quotes become stale

Core Design Contradiction:

SpeedVSLoss of time

Solution Approach 1:

The patent replaces traditional mechanical order routing systems with a machine learning-based system that uses predictive models and probabilistic algorithms to anticipate market movements and optimize order execution timing, thereby achieving faster effective processing speeds that can beat market updates

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system performs preliminary actions by using machine learning models to predict future market states and pre-position orders or adjust execution strategies before market updates occur, allowing the system to act proactively rather than reactively within the constrained time window

Inventive Principle:

Principle #10Preliminary action

2Reliability

If machine learning models are implemented to optimize order routing decisions, then the quality of execution decisions improves, but the computational complexity and processing time increase

Engineering Contradiction:

Improvequality of execution decisionsVSAvoidcomputational complexity of the system

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the machine learning system into multiple specialized predictive models, each handling specific aspects of market prediction (e.g., price movement, volume, volatility), allowing the complex problem to be divided into manageable components that can be processed more efficiently

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system dynamically adjusts model parameters and complexity based on market conditions and available time, using simpler models when speed is critical and more complex models when time permits, thereby balancing decision quality with computational burden

Inventive Principle:

Principle #35Parameter changes

3Reliability

If complex optimization algorithms are used to maximize execution quality, then the transaction outcomes improve, but the computational time required exceeds the available window before quotes become stale

Engineering Contradiction:

Improvetransaction execution qualityVSAvoidcomputational time vs. available time window

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent implements partial optimization by focusing computational resources on the most critical decision variables and using approximate algorithms that deliver sufficiently good solutions within the time constraint, rather than seeking perfect optimization that would exceed the available time window

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The system uses periodic updates and incremental optimization approaches, refining execution decisions in stages as time permits and market conditions evolve, rather than attempting to compute the complete optimal solution from scratch within the constrained time window

Inventive Principle:

Principle #19Periodic action

Data Source

PatentUS11488243B2Systems and methods for quantitative order routing

Publication Date: 2022.11.01 ROYAL BANK OF CANADA

AI summary

A smart order router for quantitative trading and order routing and corresponding methods and computer readable media are described. The smart order router includes a machine learning prediction engine configured to, responsive to a control signal received from an upstream trading engine including at least a maximum quantity value and an urgency metric, process input data sets through one or more predictive models to generate the one or more potential combinations of child orders and their associated fill probability metrics, toxicity metrics, and expected gain (loss) metrics and an order placement optimization engine configured to receive the one or more potential combinations of child orders and their associated fill probability metrics, toxicity metrics, and expected gain (loss) metrics and to identify an optimum combination of child orders that maximize an objective function.