Page 340 - Biofuels for a More Sustainable Future

296 Biofuels for a More Sustainable Future

Some studies considered the impacts of the job creation at different locations
[e.g., more develop regions versus less developed regions (Mota et al., 2015)]
or the impacts of different types of jobs created (Cambero and Sowlati, 2016).
As discussed previously, many studies used multiple objectives in opti-
mization models to design sustainable BSC. Most of them have employed
MCDA to integrate multiple objective functions into a single objective
function using assumed weighting factors (Kanzian et al., 2013; Bernardi
et al., 2013; Eskandarpour et al., 2015) or use Pareto curve to explore
trade-offs (Zhangetal.,2014; Sammons et al., 2008; Zambonietal.,
2009; Santiban ˜ez-Aguilar et al., 2016). Table 10.3 lists 61 studies of
BSC optimization reviewed and their objective functions. Among those
studies, GHG emissions reduction, job creations, and SC cost are the
mostly used indicators representing environmental, social, and economic
aspects in BSC optimization.

4.1.2 Types of BSC optimization models
The approaches to solve optimization problems vary largely according to the
different types of optimization models. The common types of BSC optimi-
zation include LP, mixed integer linear programming (MILP), MINLP,
multiobjective linear programming (MOLP), multiobjective mixed integer
linear programming (MOMILP), mixed integer quadratic programming
(MIQP), stochastic programming (SP), Fuzzy Programming (FMP), and
heuristic algorithms (HEU) (Sharma et al., 2013; Mula et al., 2010).
LP has been widely used for optimization problems as a basic approach.
LP is “concerned with problems in which a linear objective function in
terms of decision variables is to be optimized while a set of linear equa-
tions, inequalities, and sign restrictions are imposed on the decision vari-
ablesasrequirements” (Fang and Puthenpura, 1993). Jonker et al. (2016)
used the LP approach to optimize the locations and capacities of plants
given the expansion of biomass supply regions. MILP is a more common
format of BSC optimization models where binary variables are introduced
for decisions such as selecting locations, technologies, and other options. Tong
etal. (2014) employed a MILPmodeltointegratethe existing petroleum refin-
eries and biomass conversion facilities with considering the uncertainties in
production. In this model, integer variables were introduced to represent
decision selection, biorefinery property category, and other “whether or
not” variables. In some cases, MILPcan becomputationally intensive, and thus
in some studies, two-stage or multistage optimization frameworks were used
to address this challenge. Kazemzadeh and Hu (2013) adopted stochastic