IEEE Systems, Man and Cybernetics Magazine - April 2023 - 7

©SHUTTERSTOCK/BLUE PLANET STUDIO
the characteristics that influence the quality of collaborative
learning. Then, the learners' evaluation results are fed
into a specific mathematical model, and the optimal LTF
result is obtained by solving an optimization problem.
Despite the efforts of prevailing research for solving LTF
problems in different application scenarios [5], [8], [11],
[12], [13], [14], there is still a lack in learners' ability measurement
and generalized mathematical modeling methods
for solving LTF:
Previous research [5], [8] using a single real number to
measure the ability of learners and ignoring the uncertainty
of learners' ability may lead to LTF bias and consequently
reduce the quality of collaborative learning. In practice,
the real ability of one learner affected by some factors
(e.g., the material, spiritual motivation, emotional state,
knowledge level, and skill proficiency) is usually uncertain
and fluctuating. Therefore, it is helpful to evaluate learners'
real abilities accurately and objectively by introducing
a fuzzy measurement mechanism.
Existing research [11], [12], [13], [14] proposes solutions
to LTF problems in different application scenarios. However,
these problems actually share common features in
nature, and such features indicate that all LTF problems
could be essentially regarded as RBC problems. We believe
that it would provide meaningful guidance for solving LTF
problems in different scenarios if these problems could be
formalized with a generalized mathematical model.
Based on our previous experience, the E-CARGO model
is introduced to model LC-LTF problems in a universal
manner. As an architecture model for handling RBC problems,
the E-CARGO [15], [16] model uses six core concepts
to build mathematical models for RBC problems and has
been applied in many fields [17], [18]. More important, it
provides strong support for the formalization analysis of
LC-LTF problems.
Based on the characteristics analysis of two typical
types of LC-LTF problems, a universal optimization framework
for LC-LTF is proposed. This framework consists of
three modules. In the evaluation module, fuzzy theories
(e.g., the interval number [19]) are used to depict the uncertainty
of learners' ability. In the modeling module, E-CARGO
is used to formalize LC-LTF problems. In the solving
module, an optimization platform (i.e., CPLEX) is
employed to solve the problems. This article provides a
new research paradigm for LC-LTF problems.
Related Work
Table 1 provides a summary of representative literature
from multiple aspects. The column " Team Type " shows
whether a member of a team is the same as or different
from the others in important characteristics; " hybrid " indicates
that the members of a team might be quite different
in some characteristics and quite the same in other
characteristics. The column " Key Indicators " shows
April 2023 IEEE SYSTEMS, MAN, & CYBERNETICS MAGAZINE
7
IEEE Systems, Man and Cybernetics Magazine - April 2023

Table of Contents for the Digital Edition of IEEE Systems, Man and Cybernetics Magazine - April 2023
