4. Conclusions
In this paper we reviewed the most significant literature on the
use of OR methodologies to improve SWM. In particular, we
focused on strategic and tactical issues. As can be easily seen from
the tables summarizing the papers described throughout the
review, many research directions would require additional work.
In fact, the existing literature focused on strategic SWM issues is
extremely rich of sectorial contributions aimed at finding the
optimal solution for reduced problems of the more complex and
general problem described in Section 2.1. More precisely, the
literature appears poor as far as aspects like time dimension,
multi-commodity, economies of scale, and uncertainty affecting
the waste generation rates and the transportation costs over a long
time horizon. On the other hand, a huge amount of research exists
about location problems occurring in SWM systems. More efforts
are required in order to develop efficient approaches devoted to
rich SWM strategic models, in which all the aforementioned
characteristics are carefully integrated. A new challenge is represented
by studying and modeling a unified framework in which
decisions related to the collection sites, transfer stations, processing
facilities and landfills location are combined with decisions on
shipping multi-commodity waste flows on the basis of how much
profitable is to convert fractions of the waste into recycling
materials and alternative energies. Recent advances in exact and
matheuristic methodology for solving large scale MIPs represent
new powerful tools to find quickly good quality solutions of more
complex optimization problems, as those occurring in SWM
systems. More precisely, branch-and-price guided search methodology
for integer programming, and “a priori” reformulations of
the problem to be embedded into a commercial MIP solver (a
similar approach has been shown to be able to solve a nontrivial
fraction of practical lot-sizing problems by Wolsey [66] and
Gicquel et al. [67]) represent suitable approaches. Tractability of
mathematical models like those devoted to SWM systems may be
dealt with by decoupling the problem into partial subproblems to
be optimized. In such a fashion, strategic aspects may be considered
over a long-term planning horizon, while modeling the
impact of tactical decisions on the strategic solution through
tactical constraints modifying the solution obtained at each
iteration