In problems with two or more conflicting objectives, there is no
single optimum solution. There exist a number of solutions
which are optimal. Such solutions are called Pareto-optimal
solutions where no solutions dominates another solution. None
of the solutions in the Pareto-front is best with respect to all
objectives. In addition, no solution in the Pareto-front is better
than any other solution in the front with respect to all objectives.
Hence, without any additional problem-specific information
about the priorities of various objectives, all the solutions in the
Pareto-front are important. The main objective of multiobjective
optimization is to find many such solutions which
reflect the tradeoffs between the objectives.
Efficient operation of smart grids involves optimizing
energy consumption and utility which are two conflicting
objectives. Multi-objective evolutionary algorithms are
effective in obtaining a Pareto front with multiple solutions that
characterize the tradeoffs between the objectives. In our work,
we adapt a recently developed multi-objective evolutionary
algorithm (EMOCA) for handling the generation capacity
constraint while optimizing energy consumption and utility.
EMOCA is a genetic algorithm which has outperformed stateof-
the-art multi-objective optimization algorithms on several
test problems [8]. In EMOCA, each solution represents a
randomly generated energy consumption schedule for each
user. In each iteration, the solutions are modified using genetic
operators such as crossover and mutation. After several
iterations, the algorithm obtains a set of tradeoff solutions
called the Pareto-front. This set of solutions provides a wide
variety of choices for the system designer as opposed to a
combinatorial optimization approach which provides a single
solution.