This research can solve two problems: first, determine optimal delivery routes using the Clarke & Wright
algorithm, then find the optimal number of resources or servers (s) to be allocated to each transport vehicle using
total cost assessment and simulation methods.
Delivery points (customers) are grouped in clusters considering the number of customers served by each stop (N),
and with this, optimal paths are determined. Then, the attention of every stop along the routes are manually
simulated bearing in mind that each stop cluster generates a queuing system M / G / s / N, where the servers (s)
are the dispatchers and the system has N clients to attend. The analysis shows that the main problem is the
reduction of delivery time at each stop, which represents about 80% of the total time; on the other hand,
transportation itself represents about 20%. The conclusion after the economic evaluation of alternatives is that
investment in one or more dispatchers per truck helps to reduce total delivery time, overtime worked and lost
sales per workday.
|