Pathsor

Pathsor is a term used in graph theory and algorithm design to describe a family of methods for identifying and ranking multiple alternative paths between two nodes in a graph. Unlike traditional single-path shortest-path techniques, pathsor aims to produce a set of diverse, constraint-compliant paths, often ordered by a primary cost function and one or more secondary criteria such as risk, reliability, or slack time.

Origins and scope: The concept emerged in the context of multi-criteria routing and robotics planning in the 2010s as a framework for balancing efficiency with resilience. Pathsor is not a single algorithm but a design philosophy that combines path generation, dominance checks, and pruning to keep the output tractable.

Methods and characteristics: A typical pathsor pipeline starts with a base path computation (often a k-shortest paths step) and then enriches each candidate with additional attributes. It applies dominance relations to discard paths that are worse across all criteria and uses constraints to enforce resource limits or temporal windows. The result is a Pareto- or tiered-ordered set of paths suitable for decision-making.

Complexity and performance: The complexity depends on graph size, the number of paths requested, and the number of criteria. Practical implementations rely on pruning and caching to maintain scalable performance on moderate graphs.

Applications: Pathsor finds use in logistics and supply chain planning, network routing with quality-of-service considerations, and autonomous navigation where alternative routes are required for fault tolerance.

See also: k-shortest paths, multi-criteria path finding, Pareto efficiency, path diversity.

---