FDB-SDO (An improved version of Supply-Demand Optimizer)

Improving Supply-Demand-Based Optimization Algorithm with FDB Method: A Comprehensive Research on Engineering Design Problems

In FDBSDO, roulette-wheel based probabilistic version of FDB selection method was used. You can see this in the source code.

Abstract of FDBSDO: In this study, an improved version of the supply-demand-based optimization (SDO) algorithm, a recently developed meta-heuristic search method, was developed. In order to model the supply-demand processes more effectively in SDO, solution candidates guiding the search process were determined using the fitness-distance balance (FDB) method. In order to test and verify the performance of the developed FDB-based SDO algorithm, CEC 2017, a modern benchmark suite, was used. This suite has four different types and thirty unconstrained test problems. These problems are designed in 30/50/100 dimensions to test and verify the performance of the proposed algorithm in search spaces of different types and dimensions. In addition, twenty engineering design problems were used to test and verify the performance of the proposed FDBSDO variations in constrained engineering design problems. Data from both experimental studies were analyzed using non-parametric statistical test methods. According to the results of the analysis, FDBSDO variations showed superior performance compared to the base algorithm in all experimental studies, with constrained/unconstrained, unimodal/multi-modal/hybrid/composition problem types and different dimensions. The implementation of the FDB selection method eliminated the premature convergence problem of the SDO algorithm. The proposed FDBSDO algorithm has the ability to sensitively search, effectively provide diversity, and build a strong balance between exploitation-exploration.

//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////

The article introducing the FDB selection method:
https://www.sciencedirect.com/science/article/abs/pii/S0950705119305167

Up-to-date FDB based optimization algorithms (FDBSOS and FDBSFS):

1) https://www.mathworks.com/matlabcentral/fileexchange/72311-fdb-sos
2) https://www.mathworks.com/matlabcentral/fileexchange/84405-fdb-sfs