Rational and self-adaptive evolutionary extreme learning machine for electricity price forecast

Electricity price forecast is of great importance to electricity market participants. Moreover, various prediction approaches based on extreme learning machine (ELM) have been identified as effective on normal decision space. Especially, evolutionary extreme learning machine (E-ELM) may obtain better solution quality. However, in high dimensional space, E-ELM is time-consuming because it is difficult to converge into optimal region when just relied on stochastic searching approaches. In addition, due to the complex functional relationship is often complicated, the objective function of E-ELM seems hard to be mined directly for obtaining useful mathematical information to guide the optimum exploring. This paper proposes a new differential evolution (DE)-like algorithm to enhance E-ELM for more accurate and reliable prediction of electricity price. The approximation model for producing DE-like trail vector is the key mechanism, which may use simpler mathematical mapping to replace the original yet complicated functional relationship within a small region. Thus, the evolutionary procedure frequently guided by rational searching directions may make the E-ELM more robust and faster than supported only by those stochastic methods. Several benchmarks are applied to test the performances of the proposed algorithm and the experimental results have shown that the new method can improve the performance of E-ELM.