Cost Optimization by Method of Allocating Software Component Units to Electronic Control Units for Model-Driven Designs

Abstract

Technological advancement in the semiconductor industry and applications has caused a tremendous growth in vehicular electronics, one of the major hurdles due to this explosion is the increase in wiring beyond proportions. Due to the increase in wiring, mileage, stability and reliability of the automotive is affected. At this critical juncture, model based designs allows a software implementation of the automotive and hence multiple architectures can be designed and evaluated at a faster rate without the necessity of building an actual automotive. A binary quadratic programming model (BQP) is designed to optimize the wiring and also consider the total cost of the electronics, so that an optimal minimal cost solution can be achieved with respect to the electronics of the automotive. The solution of BQP model will result in the assignment of software components (software tasks) to electronic control units (processors) based on various constraints like memory, performance e.t.c. and to achieve the wiring harness. A software package which will setup the environment, solve the binary quadratic programming model to optimality and provide solution sets will be presented here. The thesis work will provide benchmarks for various problems and a variant of an actual automobile. These benchmarks range from simple examples to futuristic vehicles and provide an insight into the optimization achieved from a system perspective. A simple example is used to illustrate the parameters of the BQP model and results are presented.