Optimized Selection of Reliable and Cost-Effective Cyber-Physical System Architectures
Nikunj Bajaj1,a, Pierluigi Nuzzo1,b, Michael Masin2 and Alberto Sangiovanni-Vincentelli1,c
1EECS Department, University of California at Berkeley, USA.
2IBM Haifa Research Lab, Haifa, Israel.
We address the problem of synthesizing safetycritical cyber-physical system architectures to minimize a cost function while guaranteeing the desired reliability. We cast the problem as an integer linear program on a reconfigurable graph which models the architecture. Since generating symbolic probability constraints by exhaustive enumeration of failure cases on all possible graph configurations takes exponential time, we propose two algorithms to decrease the problem complexity, i.e. Integer-Linear Programming Modulo Reliability (ILP-MR) and Integer-Linear Programming with Approximate Reliability (ILP-AR). We compare the two approaches and demonstrate their effectiveness on the design of aircraft electric power system architectures.
Full Text (PDF)