This paper presents a two-level approach to designing and analyzing se
mi-automated mail processing facilities. The strategic issues related
to equipment selection and capacity expansion are examined at the uppe
r level with the help of a large scale mixed integer linear program. T
he general formulation is characteristic of a multistage, multicommodi
ty network, and contains over 7500 variables and 6500 constraints. Sol
utions provide input to a SLAM-based simulation model that is used to
investigate operational issues related to resource scheduling, service
standards, and growth in mail volume. The Northern Virginia Area faci
lity serves as the baseline for the study. Results are presented for a
proposed bodying a host of new concepts and equipment. The first is a
n upgraded multi-line optical character reader (MLOCR) which is capabl
e of scanning the entire address field, looking up a corresponding 11-
digit ZIP code in a national directory, and bar-coding the envelope ac
cordingly. The second is a remote video encoding system (RVES) which w
ill perform the same functions as the MLOCR, but with an operator at a
console to process handwritten addresses. with this new equipment, it
will be possible to sequence almost 80 percent of the letter mail to
the order in which it is delivered by the carrier. This is expected to
reduce the amount of in-office activity by up to 45 percent, and save
over $6 billion a year when implemented nationwide.