Software-to-hardware allocation plays an important role in the development of resource-constrained automotive embedded systems that are required to meet timing, reliability and power requirements. This paper proposes an Integer Linear Programming optimization approach for the allocation of faulttolerant embedded software applications that are developed using the AUTOSAR standard. The allocation takes into account the timing and reliability requirements of the multirate software applications and the heterogeneity of their execution platforms. The optimization objective is to minimize the total power consumption of the applications that are distributed over multiple computing units. The proposed approach is evaluated using a range of different software applications from the automotive domain, which are generated using the real-world automotive benchmark. The evaluation results indicate that our proposed allocation approach is effective while meeting the timing, reliability, and power requirements of the considered automotive software applications.