Acquisition and Active Navigation of Knowledge Particles throughout Product Variation Design Process

The variation design of complex products has such features as multivariate association, weak theory coupling and implicit knowledge iteration. However, present CAD soft wares are still restricted to making decisions only according to current design status in dynamic navigation which leads to the huge drain of the knowledge hidden in design process. In this paper, a method of acquisition and active navigation of knowledge particles throughout product variation design process is put forward. The multi-objective decision information model of the variation design is established via the definition of condition attribute set and decision attribute set in finite universe. The addition and retrieval of the variation semantics is achieved through bidirectional association between the transplantable structures and variation design semantics. The mapping relationships between the topology lapping geometry elements set and constraint relations set family is built by means of geometry feature analysis. The acquisition of knowledge particles is implemented by attribute reduction based on rough set theory to make multi-objective decision of variation design. The topology lapping status of transplantable substructures is known from DOF reduction. The active navigation of knowledge particles is realized through embedded event-condition-action（ECA） rules. The independent prototype system taking Alan, Charles, Ian＇s system（ACIS） as kernel has been developed to verify the proposed method by applying variation design of complex mechanical products. The test results demonstrate that the navigation decision basis can be successfully extended from static isolated design status to dynamic continuous design process so that it more flexibly adapts to the different designers and various variation design steps. It is of profound significance for enhancing system intelligence as well as improving design quality and efficiency.

Efficient SRAM yield optimization with mixture surrogate modeling

Largely repeated cells such as SRAM cells usually require extremely low failure-rate to ensure a mod- erate chi yield. Though fast Monte Carlo methods such as importance sampling and its variants can be used for yield estimation, they are still very expensive if one needs to perform optimization based on such estimations. Typ- ically the process of yield calculation requires a lot of SPICE simulation. The circuit SPICE simulation analysis accounted for the largest proportion of time in the process yield calculation. In the paper, a new method is proposed to address this issue. The key idea is to establish an efficient mixture surrogate model. The surrogate model is based on the design variables and process variables. This model construction method is based on the SPICE simulation to get a certain amount of sample points, these points are trained for mixture surrogate model by the lasso algorithm. Experimental results show that the proposed model is able to calculate accurate yield successfully and it brings significant speed ups to the calculation of failure rate. Based on the model, we made a further accelerated algo- rithm to further enhance the speed of the yield calculation. It is suitable for high-dimensional process variables and multi-performance applications.