Constraint management for collaborative electronic design

Carballo, Juan Antonio; Stephen, W

doi:10.1145/309847.309992

“…The DCM then runs a constraint propagation algorithm to compute infeasible property values and the status of all constraints. Constraint evaluation details are delegated to constraint-based systems and CAD tools [3]. The result is sent back to the DPM, which properly updates C n+1 and the status of design problems.…”

Section: The Adpm Design Process Modelmentioning

confidence: 99%

“…We have implemented constraint-based heuristic support in the Minerva III design process manager [3]. We describe this support with an example: the team-based design of a MEMS-based wireless receiver front-end subject to gain, power, bandwidth, and frequency precision constraints.…”

Section: Constraint-based Heuristics In Minerva IIImentioning

confidence: 99%

“…ADPM is based on a design process modeling framework that is built on previous work [3,8,10] and emphasizes the role of constraints. In this framework, a design is characterized by a set of variables called properties.…”

Section: Introductionmentioning

confidence: 99%

“…A key challenge is the complexity of the required constraint management infrastructure. A methodology called CCM [3] was introduced whereby this infrastructure is developed on the basis of constraint-based systems [1,9] and CAD tools. We leverage this work by computing constraint-related information using CCM's constraint generation and propagation techniques, and then "mining" the results into data that directly supports search heuristics (e.g., the number of violations related to each design variable).…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Application of constraint-based heuristics in collaborative design

Carballo

¹

,

Stephen

²

2001

Proceedings of the 38th Conference on Design Automation - DAC '01

Self Cite

3

0

View full text Add to dashboard Cite

Significant acceleration of today's complex collaborative design processes can be achieved if team members are able to apply search heuristics that consider the simultaneous effect of all design constraints. We present the Active approach to Design Process Management (ADPM), whereby designers receive constraint-based feedback that enables them to apply these search heuristics effectively. To evaluate ADPM, we developed a design process evaluation environment called TeamSim. Evaluation results suggest that ADPM can reduce costly design iterations at the expense of extra, less costly, verification tool executions. INTRODUCTIONComplex electronic designs are subject to ever tighter time-tomarket requirements and thus involve ever larger teams, where multiple subsystems are developed in parallel by different groups. Unfortunately, this concurrent design often results in the late detection of conflicts involving multiple subsystems. These conflicts tend to be found upon system integration and thus are very costly to resolve. If one views the design as a group of variables subject to a set of constraints, conflicts can then be seen as violations of constraints. Late conflict detection occurs because, until system integration, each group of designers typically considers only a subset of all constraints relating their subsystem to other subsystems. We can substantially reduce costly rework by aiding consideration of the simultaneous effect of all constraints. The amount and variety of constraints makes computer support for this task essential. For this support to be most powerful, it must give designers direct instructions or "clues" to improve their design space search throughout the design process. This paper presents Active Design Process Management (ADPM), a state-based design process model whereby team members receive constraint-based feedback on their operations and use it to apply design space search heuristics effectively. This guidance reduces and helps resolve conflicts. We also present TeamSim, a design process evaluation environment developed on top of the Minerva III design process manager [3] to evaluate ADPM.Design can be viewed as a search process in a design space restricted by constraints. Constraint-based search heuristics can substantially improve search algorithms [2,6,9] and thus may significantly accelerate design convergence. Several types of constraint-based information can help effectively apply these heuristics, including:• Infeasible design subspaces. The design process may be accelerated by focusing first on areas of the design space that have the smallest subspaces not found to be infeasible.• Strongly constrained subspaces. Another heuristic is to focus first on design subspaces affected by the most constraints.• Efficient conflict resolution strategies. Design convergence may also be accelerated by (a) making use of trade-offs produced by constraint margins to fix violations and (b) executing design operations that will fix many violations at a time. While heuristics are often used...

show abstract

“…This guidance reduces and helps resolve conflicts. We also present TeamSim, a design process evaluation environment developed on top of the Minerva III design process manager [3] to evaluate ADPM.…”

Section: Introductionmentioning

confidence: 99%

Application of constraint-based heuristics in collaborative design

Carballo¹,

Director²

Proceedings of the 38th Design Automation Conference (IEEE Cat. No.01CH37232)

View full text Add to dashboard Cite

Significant acceleration of today's complex collaborative design processes can be achieved if team members are able to apply search heuristics that consider the simultaneous effect of all design constraints. We present the Active approach to Design Process Management (ADPM), whereby designers receive constraint-based feedback that enables them to apply these search heuristics effectively. To evaluate ADPM, we developed a design process evaluation environment called TeamSim. Evaluation results suggest that ADPM can reduce costly design iterations at the expense of extra, less costly, verification tool executions. INTRODUCTIONComplex electronic designs are subject to ever tighter time-tomarket requirements and thus involve ever larger teams, where multiple subsystems are developed in parallel by different groups. Unfortunately, this concurrent design often results in the late detection of conflicts involving multiple subsystems. These conflicts tend to be found upon system integration and thus are very costly to resolve. If one views the design as a group of variables subject to a set of constraints, conflicts can then be seen as violations of constraints. Late conflict detection occurs because, until system integration, each group of designers typically considers only a subset of all constraints relating their subsystem to other subsystems. We can substantially reduce costly rework by aiding consideration of the simultaneous effect of all constraints. The amount and variety of constraints makes computer support for this task essential. For this support to be most powerful, it must give designers direct instructions or "clues" to improve their design space search throughout the design process. This paper presents Active Design Process Management (ADPM), a state-based design process model whereby team members receive constraint-based feedback on their operations and use it to apply design space search heuristics effectively. This guidance reduces and helps resolve conflicts. We also present TeamSim, a design process evaluation environment developed on top of the Minerva III design process manager [3] to evaluate ADPM.Design can be viewed as a search process in a design space restricted by constraints. Constraint-based search heuristics can substantially improve search algorithms [2,6,9] and thus may significantly accelerate design convergence. Several types of constraint-based information can help effectively apply these heuristics, including:• Infeasible design subspaces. The design process may be accelerated by focusing first on areas of the design space that have the smallest subspaces not found to be infeasible.• Strongly constrained subspaces. Another heuristic is to focus first on design subspaces affected by the most constraints.• Efficient conflict resolution strategies. Design convergence may also be accelerated by (a) making use of trade-offs produced by constraint margins to fix violations and (b) executing design operations that will fix many violations at a time. While heuristics are often used...

show abstract