Inverse Design Fabricability via Random Perturbations

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current design techniques for optical and electromagnetic devices often rely on guesswork and do not adequately consider the fabricability of generated designs, leading to inefficiencies as device complexity increases and feature sizes decrease.

Innovation Solution

A method involving a computing system that receives a design specification, optimizes an initial segmented design, perturbs it to create a perturbed design, and further optimizes this to produce a fabricable segmented design, using simulations and adjoint simulations to iteratively update structural parameters and avoid local minima in loss functions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If simple guess and check method is used for device design, then design process is simple, but device performance optimization is insufficient

Engineering Contradiction:

Improvedesign process complexityVSAvoiddevice performance optimization

Core Design Contradiction:

Device complexityVSManufacturing precision

Solution Approach 1:

The patent applies parameter changes by systematically varying design parameters through optimization algorithms rather than simple guess and check. The method transforms the design space into a parameterized representation and uses gradient-based optimization to efficiently explore and optimize these parameters, achieving superior device performance while managing design complexity through structured parameter exploration.

Inventive Principle:

Principle #35Parameter changes

2Productivity

If design techniques generate device designs without considering fabricability, then design generation is faster, but manufacturing feasibility is compromised

Engineering Contradiction:

Improvedesign generation speedVSAvoidfabricability

Core Design Contradiction:

ProductivityVSEase of manufacture

Solution Approach 1:

The patent applies preliminary action by incorporating fabricability constraints into the design generation process itself, rather than checking for manufacturability after design completion. The method integrates manufacturing constraints as boundary conditions during the optimization process, ensuring that generated designs inherently satisfy fabricability requirements while maintaining design generation efficiency through guided exploration of the design space.

Inventive Principle:

Principle #10Preliminary action

3Loss of time

If optimization process does not use perturbations, then convergence is faster, but likelihood of escaping local minima is reduced

Engineering Contradiction:

Improveoptimization convergence timeVSAvoidglobal optimality

Core Design Contradiction:

Loss of timeVSReliability

Solution Approach 1:

The patent applies periodic action by implementing a multi-stage optimization process that alternates between gradient-based optimization phases and perturbation phases. The method periodically introduces controlled perturbations to the design parameters at strategic points during optimization, allowing the system to escape local minima while maintaining overall convergence efficiency through the structured alternation between exploitation and exploration phases.

Inventive Principle:

Principle #19Periodic action

Data Source

PatentUS11630932B2Techniques for using random perturbations during an inverse design process to obtain fabricable designs

Publication Date: 2023.04.18 X DEVELOPMENT LLC

AI summary

A method of creating a fabricable segmented design for a physical device is provided. A computing system receives a design specification. The computing system optimizes an initial segmented design based on the design specification to create an improved segmented design. The computing system perturbs the improved segmented design to create a perturbed segmented design. The computing system optimizes the perturbed segmented design to create a second improved segmented design.