Online Inverse Optimization for Time-Varying Decision Objectives

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing inverse optimization techniques struggle to handle time-varying optimization problems and recover both objectives and constraints in an online manner, particularly in complex scenarios like employee scheduling, where data quality issues and agent decisions are involved.

Innovation Solution

An online probabilistic inverse optimization system that infers objectives and constraints using machine learning techniques, treating the problem as a maximum likelihood estimation, allowing for flexible adaptation to changing data and agent decisions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If standard inverse optimization techniques are used, then objectives can be recovered from agent decisions, but the techniques cannot handle time-varying optimization problems

Engineering Contradiction:

Improveability to handle time-varying problemsVSAvoidaccuracy of objective recovery

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The patent transforms the static inverse optimization problem into a dynamic online learning framework where objectives are continuously updated as new agent decisions arrive. The system adapts to time-varying problems by processing sequential data and updating objective estimates in real-time, making the recovery process dynamic rather than static.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system implements feedback by using observed agent decisions to continuously refine and update the recovered objectives. Each new decision provides feedback that adjusts the objective estimates, enabling the system to adapt to changing conditions while maintaining accuracy through iterative refinement.

Inventive Principle:

Principle #23Feedback

2Loss of information

If multiple objectives and constraints are considered in inverse optimization, then more comprehensive agent behavior is captured, but the complexity of the optimization problem increases

Engineering Contradiction:

Improvecompleteness of recovered objectivesVSAvoidcomplexity of optimization problem

Core Design Contradiction:

Loss of informationVSDevice complexity

Solution Approach 1:

The patent segments the complex multi-objective optimization problem into simpler component objectives that can be recovered and evaluated separately. By breaking down the overall objective function into multiple individual objectives, the system can handle complexity through modular processing while maintaining comprehensive coverage of agent behavior.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system changes parameters by representing multiple objectives and constraints as learnable parameters in a probabilistic framework. This allows the complex problem to be transformed into a parameter estimation task that can be solved through online learning, reducing computational complexity while preserving information about multiple objectives.

Inventive Principle:

Principle #35Parameter changes

3Adaptability or versatility

If online learning is used to adapt to changing data, then the system can handle dynamic environments, but data quality issues and noise affect the learning accuracy

Engineering Contradiction:

Improvereal-time adaptation capabilityVSAvoidaccuracy of objective inference

Core Design Contradiction:

Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent converts the harmful effect of noise and data quality issues into a benefit by using a probabilistic framework that explicitly models uncertainty. Rather than treating noise as an error to be eliminated, the system incorporates it into the learning process, where uncertain observations naturally lead to uncertain objective estimates that are continuously refined as more data arrives.

Inventive Principle:

Principle #22Blessing in disguise (Convert harm into benefit)

Solution Approach 2:

The system provides beforehand cushioning by using a probabilistic framework that anticipates and buffers against data quality issues. The uncertainty modeling acts as a cushion that prevents noise from causing large deviations in objective recovery, allowing the system to maintain stability and accuracy even when individual observations are noisy or unreliable.

Inventive Principle:

Principle #11Beforehand cushioning (Prior cushioning)

Data Source

PatentUS20260080270A1Online probabilistic inverse optimization system for decision making support, online probabilistic inverse optimization method for decision making support, and online probabilistic inverse optimization program for decision making support

Publication Date: 2026.03.19 NEC CORP

AI summary

An online probabilistic inverse optimization system 10 is proposed for inferring objectives and constraints in an online fashion from changing problem data and corresponding agent decisions. The online probabilistic inverse optimization system 10 includes: a computing unit 11 which computes optimal solutions or decisions based on the forward optimization problem using the problem data that may include objectives, constraints and parameters; and a solving unit 12 which solves the inverse optimization problem using the agent decisions.