1: \begin{abstract}

2:   We present a systematic approach to the optimal placement of finitely many sensors in

3:   order to infer a finite-dimensional parameter from point evaluations of the solution of

4:   an associated parameter-dependent elliptic PDE. The quality of the corresponding least

5:   squares estimator is quantified by properties of the asymptotic covariance

6:   matrix depending on the distribution of the measurement sensors. We formulate a design

7:   problem where we minimize functionals related to the

8:   size of the corresponding confidence regions with respect to the position and number

9:   of pointwise measurements. The measurement setup is modeled by a positive Borel measure

10:   on the spatial experimental domain resulting in a convex optimization problem.

11:   For the algorithmic solution a class of accelerated conditional gradient methods in

12:   measure space is derived, which exploits the structural properties of the design problem to ensure convergence

13:   towards sparse solutions. Convergence properties are presented 

14:   %and our findings are compared to previous results. A

15:   %variational discretization of the continuous problem based on finite elements is investigated

16:   and the presented results are illustrated by numerical experiments.

17: \keywords{ PDE-constrained inverse problems \and Optimal sensor placement \and Sparsity \and Conditional gradient methods }

18: %\PACS{PACS code1 \and PACS code2 \and more}

19:  \subclass{35R30 \and 49K20 \and 62K05 \and 65K05 \and 49M25 }

20: \end{abstract}

21: