Random Solution Sampling

Concept

Random Solution Sampling is a requirement in constraint-based stimuli generation where a single design model and test template define one Soft-CSP, but the solver is expected to produce many different tests that are as uniformly distributed as possible over the conforming solution space.

First seen 5/26/2026

Last seen 5/26/2026

Evidence 1 chunks

Wiki v1

WIKI

Overview

Random Solution Sampling is described as a distinguishing requirement of constraint satisfaction problems arising in stimuli generation. In this setting, a design model together with a test template defines a single Soft-CSP, but the intended use is to obtain many different tests from that same template. The desired behavior is for those tests to be distributed as uniformly as possible among all tests that conform to the template.

Operationally, this means the solver should reach a significantly different solution each time it is run on the same CSP.

READ FULL ARTICLE →

NEIGHBORHOOD

No graph connections found for this entity yet. It may appear in future ingestion runs.

explore full graph →

RELATIONSHIPS

1 connections

Genesys PE ← uses 100% 1e

Genesys PE randomizes decisions in the MAC search to achieve disperse solutions from the same template.

LINKED ENTITIES

2 links

Soft-CSP uses_problem_formulation The evidence states that a design model and test template define a single Soft-CSP and that CSP technology supports prioritization using a generic Soft-CSP algorithm.

Constraint Hierarchy interacts_with The evidence states that constraint hierarchy competes with the requirement for uniformly distributed solutions and describes a local-metric workaround.

CITATIONS

8 sources

8 citations — click to expand

[1] In stimuli generation, a design model and test template define a single Soft-CSP, but the solver is expected to produce many different tests that are as uniformly distributed as possible over conforming tests. [PDF] Constraint-Based Random Stimuli Generation for Hardware ... - AAAI

[2] Random solution sampling requires the solver to reach a significantly different solution each time it is run on the same CSP. [PDF] Constraint-Based Random Stimuli Generation for Hardware ... - AAAI

[3] The cited approach randomizes all decisions in the MAC search path and reports that this yields reasonably disperse solutions. [PDF] Constraint-Based Random Stimuli Generation for Hardware ... - AAAI

[4] Randomizing all MAC decisions prevents the use of variable/value ordering heuristics and creates a major deviation from regular MAC-based techniques. [PDF] Constraint-Based Random Stimuli Generation for Hardware ... - AAAI

[5] CSP is used because it is declarative and because Soft-CSP methods can account for prioritization of expert-knowledge rules. [PDF] Constraint-Based Random Stimuli Generation for Hardware ... - AAAI

[6] The specialized solver framework uses maintain-arc-consistency, and stochastic search is used in rare cases where constraint propagation is computationally hard. [PDF] Constraint-Based Random Stimuli Generation for Hardware ... - AAAI

[7] Constraint hierarchies can conflict with uniformly distributed random solutions, because maximizing soft-constraint satisfaction may force the same unique assignment each run. [PDF] Constraint-Based Random Stimuli Generation for Hardware ... - AAAI

[8] The cited workaround uses a local metric: extend partial solutions to satisfy additional soft constraints when possible, but drop the requirement to maximize soft-constraint satisfaction over the entire search space. [PDF] Constraint-Based Random Stimuli Generation for Hardware ... - AAAI