I. Introduction
In real life, many applications can be modeled as multiobjective combinatorial optimization (MOCO) [1] [2]. For example, in transportation, minimizing logistics time (cost) and maximizing logistics security are a pair of objectives that need to be considered jointly. The traveling salesman problem (TSP) [3] is a subfield of combinatorial optimization (CO). The bi-objective traveling salesmen (agents) problem (BOTSP) belonged to MOCO is the extension of TSP, where two contradictory optimization goals need to be jointly optimized. Compared with normative TSP, BOTSP is more difficult to optimize. The difficulties are mainly reflected in the fact that the weight coefficients of sub-objective in the joint objective vary on different situations. It is expected to find a set of compromise solutions called Pareto optimal solutions for BOTSP.