Scheduling of Solid Waste Collection Routes

ABSTRACT ( . 1, Routing of solid waste collection vehicles in Nigeria poses a challenging task I because ol' attitudinal and liapliazard infi-a structural problems to contend with. 1, New decision procedures for solid waste collection problem are introduced in this study. 'The problem objective is to minimise the overall cost, which is essentially I based on tlie distance travelled by vehicle. The study proposes heuristic methods to generate a feasible solution to an extended Capacitated Chinese Postman I Problem (CCPI-') on the undirected network, inspired by the refuse collection proble~iis in Abuja and Onitsha. CCPP is known to be NP hard. The heuristic procedures consist of "route first, cluster second" and "cluster first, route scconcl". 'l'he compi~tational experience wit11 tlie heuristics in Abuja and Onitsha is presented. 'The two techniques were compared and with the existing schedule with respect to cost, efficiency, and distance travelled. The results demonstrated the practical applicability of the proposed ~nethodology. A cost model was developed to compare the quality of solution derived. In Okpolto, the use of *container trucks (16m3) gave sooute length of 36.03 kni, car loading compaction buck (7.5 in3) gave 53.39 knl, and side loader (5 m') gave 78.66 km. Route length increased with decreased vehicle size. The adoption of the proposed heuristics in Onitsha resulted in reduction of the ni~niber 01' existing vehicles by tires, 81 44,000 (or 7.65%) in rel'use collection cost and 28.17 Itni (or 6.03%) in vehicle distance travel led per day. In addition, in Abuja, the heuristics produced routes, wl. lich could save about 19.08 km travel per day and $4 4,197.6 (or 2 1.09%) cf collection cost per day. Finally, the heuristics increased the efficiency of refuse collection in Abi!ja from 86% to 98% and in Onitsha from 75% lo 95%. I The results revealed good performance of the proposed heuristic methods, I which will find useful applications in other areas of' vehicle scheduling.