The Traveling Analyst Problem (TAP) [2] can be defined as follows : let Q be a set of n
database queries labeled q 1 through q n , defined by an interstingness score v i and an execution
time t i . Besides between q i and q j is's defined a distance c i,j representing the suitability of
finding q j right after q i . The goal is finding the most interesting sequence such that the overall
time and distance are bounded by ϵ t and ϵ d respectively.
The TAP can be seen as an extension to the orienteering problem (OP) [4] which originates
from the database community where it represents a data exploration problem. The TAP is
NP-hard and we focus on its heuristic solution. Matheurisitcs are known for producing near
optimal solutions in a reasonable amount of time by exploiting mixed integer programming
solvers. [3]. To the best of our knowledge there are only two previous works describing the use
of matheuristics for solving related problems : [1] focuses on the team orienteering problem
and [5] on time dependent profits. These matheurisitc have shown great efficiency compared
to state of the art algorithms. We propose the first matheuristic to solve the TAP and show
its efficiency by means of computational experiments.