Abstract
Computing cost-optimal paths in network data is an important task in many application areas like transportation networks, computer networks, or social graphs. In many cases, the cost of an edge can be described by various cost criteria. For example, in a road network possible cost criteria are distance, time, ascent, energy consumption or toll fees. In such a multicriteria network, path optimality can be defined in various ways. In particular, optimality might be defined as a combination of the given cost factors. To avoid finding a suitable combination function, methods like path skyline queries return all potentially optimal paths. To compute alternative paths in larger networks, most efficient algorithms rely on lower bound cost estimations to approximate the remaining costs from an arbitrary node to the specified target. In this paper, we introduce ParetoPrep, a new method for efficient lower bound computation which can be used as a preprocessing step in multiple algorithms for computing path alternatives. ParetoPrep requires less time and visits less nodes in the network than state-of-the-art preprocessing steps. Our experiments show that path skyline and linear path skyline computation can be significantly accelareted by ParetoPrep.