Note: Demonstrating analytics in a low-tech context–truck-routing for solid-waste collection in an Indian metropolis.

• Introducing analytics in a low-tech waste collection environment. • Demonstrating a range of analytics-based solutions for truck routing. • Demonstrating solution time (heuristic) vs. solution quality (optimality) • Tactical evaluation of routes with nearest neighbor and MILP-based VRP. • Strategic evaluation of existing partition of nodes with MST and K-medoids. This paper describes an approach to introducing analytics through various algorithms and applications to users in a low-tech environment as a first step toward understanding such a context. The South Delhi Municipal Corporation (SDMC) of New Delhi, India, have partitioned their collection points into "wards" or clusters, each served by a dedicated truck depot and manually routing trucks for solid waste collection within each ward, with the waste from all wards going to a single landfill. To demonstrate analytics in tactical planning, we implemented the nearest neighbor algorithm mimicking the manual process to provide the baseline cost. Thus, we presented two very different vehicle routing algorithms: (1) a simple but fast revised nearest neighbor algorithm that decreased the baseline total routing cost by 1.57 % and (2) an optimal but time-intensive algorithm using a mixed-integer-linear programming model, which decreased the total cost by 4.05 %. To demonstrate strategic planning, we tested the efficacy of the cluster structure of collection points by comparing its total routing cost (using the revised nearest neighbor algorithm) to that of other partitions obtained with Minimum Spanning Tree (MST) and K-medoids clustering. The existing wards provided a lower waste pickup cost than the alternative clusters we created, showing SDMC that their existing ward structure was sound. [ABSTRACT FROM AUTHOR]