A multi-objective sustainable multipath delivery problem in hilly regions with customer-satisfaction using TLBO

Logistic delivery through road contributes substantial carbon emission (CE). In business, timely goods delivery i.e. customer satisfaction, is important. With these facts, a sustainable multi-objective 3D delivery problem with customer satisfaction (SMO3DDPwCS) in a hilly region (HR) is developed to minimize total CE and customer dissatisfaction (CDS) simultaneously. Here, one supplier’s vehicle starts from the depot with goods equal to retailers’ demands, distributes among the retailers as per their orders within their preferred times, and comes back. The retailers’ shops and depot are connected through multiple hilly tracks, which have up and down slopes and are susceptible to landslide. The cautious driving through these tracks produces extra CE and CDS. The SMO3DDPwCS is solved by a modified MOTLBO (mMOTLBO) algorithm. This algorithm incorporates self-learning concepts after both the teaching and learning phases, introduces innovative upgrading strategies, and employs a group-based learning approach. Some statistical tests are performed using mMOTLBO on the standard TSPLIB instances. The efficiency of mMOTLBO is established against NSGA-II and MOEA/D. Multiple solutions in Pareto front are ranked using TOPSIS. Some managerial decisions are drawn. The optimum routing plan for SMO3DDPwCS in a hilly region is presented and gives better results (31% total CE and 8% total CDS) than the single path formulation. mMOTLBO showed superiority over other algorithms in most cases concerning the Pareto front for the objectives. On the benchmark instances, mMOTLBO demonstrated its superiority by outperforming NSGA-II and MOEA/D, showing improvements of 0.11 in IGD and 4.12 in GD. • Sustainable multi-objective multipath delivery models in hilly region (HR) developed. • Road gradient, landslide, recovery time in HR, delayed delivery time, cost taken. • Carbon emission and customers’ dissatisfaction minimized by modified TLBO(mMTLBO). • Statistical tests on TSPLIB problems by mTLBO, efficiency against NSGA-II and MOEA/D. • Real life routing problem solved by mTLBO, Pareto fronts presented, ranked by TOPSIS.