Theory of second optimization for scan experiment

TL;DR

This paper develops a second optimization theory for scan experiments in high energy physics, determining optimal energy points and luminosity allocation for parameter estimation, applicable to any scan design.

Contribution

It introduces a novel second optimization framework that analytically and sampling-based determines optimal scan points and luminosity distribution for multi-parameter experiments.

Findings

Optimal number of energy points equals the number of parameters.

Single parameter scans suffice to find each optimal energy point.

Luminosity can be allocated analytically based on parameter importance.

Abstract

In many high energy experiments, the physics quantities are obtained by measuring the cross sections at a few energy points over an energy region. This was referred to as scan experiment. The optimal design of the scan experiment (how many energy points, what the energies are, and what is the luminosity at each energy point) is of great significance both for scientific research and from economical viewpoint. Two approaches, one has recourse to the sampling technique and the other resorts to the analytical proof, are adopted to figure out the optimized scan scheme for the relevant parameters. The final results indicate that for $n$ parameters scan experiment, $n$ energy points are necessary and sufficient for optimal determination of these $n$ parameters; each optimal position can be acquired by single parameter scan (sampling method), or by analysis of auxiliary function (analytic method); the luminosity allocation among the points can be determined analytically with respect to the relative importance between parameters. By virtue of the second optimization theory established in this paper, it is feasible to accommodate the perfectly optimal scheme for any scan experiment.