Figure 4
A line graph compares the average computation time per instance for nominal-operation instances between two planners, OPTIC and POPF. The x-axis represents the nominal-operation instance scale, labeled as S1, S2, M1, M2, L1, L2, L3, VL1, VL2, and VL3. The y-axis represents the average computation time in seconds, ranging from 0 to 100. The graph shows two data lines: one for OPTIC, represented by a solid line with square markers, and one for POPF, represented by a dashed line with circular markers. Both planners scale smoothly and monotonically. POPF is faster up to scale M1, after which OPTIC becomes faster, with the gap widening significantly at VL3, where OPTIC takes 64.6 seconds and POPF takes 95.3 seconds. All values are approximated.Average computation time per instance for nominal-operation instances. Both planners scale smoothly and monotonically, POPF is faster up to scale M1 and OPTIC from M2 onward, with the gap widening to 1.5× at VL3 (64.6 s against 95.3 s)