Stochastic skyline route planning under time-varying uncertainty

Searching Trajectories by Regions of Interest

“…The weight W (e) of an edge e represents the corresponding road segment's length or some other relevant property such as its fuel consumption [10], [22] or travel time [9], which can be acquired by mining historic traffic data. When weights model aspects (e.g., fuel consumption and travel time), the lower bound of network distance may not be the corresponding Euclidean distance; therefore, spatial indexes such as the R-tree [11] are ineffective.…”

“…Trajectories in T f are sorted according to the value of C sd (C, τ ).ub and are refined from the maximum to the minimum. Once max τ ∈Tr {C sd (C, τ )} ≥ max τ ∈Tu {C sd (C, τ ).ub}, the refinement terminates, and the trajectory with the maximum spatial-density correlation is returned (lines [19][20][21][22].…”

Collective Travel Planning in Spatial Networks

“…The weight W (e) of an edge e represents the corresponding road segment's length or some other relevant property such as its travel time [9] or fuel consumption [12], [28], which may be obtained from historical traffic data. Given two vertices p a and p b in a spatial network, the network shortest path between them (i.e., a sequence of edges linking p a and p b where the accumulated weight is minimal) is denoted by SP (p a , p b ), and its length is denoted by sd(p a , p b ).…”

Towards Total Traffic Awareness

“…Eco-routing needs historical knowledge to construct time-varying eco-weights that describe emissions on road segments across time and may also need instant and future knowledge to update eco-weights. Concerns for travel time and distance may also be integrated into eco-routing, yielding multi-criteria routing [20]. Next, continuousrouting utilizes current and future knowledge to provide up-to-date routes, e.g., the fastest route, from a driver's current location to the driver's destination as traffic conditions change.…”