<div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr" class="gmail-m_-7094607949299735570gmail_signature"><div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr"><div><div>Hello!<br></div></div><div><br></div><div>Currently DE implementation in NLPSover has a coefficient called FACTOR:<br></div><div><br></div><div><a href="https://github.com/LibreOffice/core/blob/f4ba484183a1e7b9824f10580d633466c266828f/nlpsolver/ThirdParty/EvolutionarySolver/src/net/adaptivebox/deps/behavior/DEGTBehavior.java#L41" target="_blank">https://github.com/LibreOffice/core/blob/f4ba484183a1e7b9824f10580d633466c266828f/nlpsolver/ThirdParty/EvolutionarySolver/src/net/adaptivebox/deps/behavior/DEGTBehavior.java#L41</a><br></div><div><br></div><div><a href="https://github.com/LibreOffice/core/blob/330df37c7e2af0564bcd2de1f171bed4befcc074/nlpsolver/src/com/sun/star/comp/Calc/NLPSolver/DEPSSolverImpl.java#L129">https://github.com/LibreOffice/core/blob/330df37c7e2af0564bcd2de1f171bed4befcc074/nlpsolver/src/com/sun/star/comp/Calc/NLPSolver/DEPSSolverImpl.java#L129</a><br></div><div><br></div><div>There are some researches about multiple values of the scaling factor: </div><div><br></div><div><a href="https://www.sciencedirect.com/science/article/pii/S0020025511000909">https://www.sciencedirect.com/science/article/pii/S0020025511000909</a><br></div><div><br></div><div>My idea is FACTOR to be replaced with two coefficients:<br><br>MIN_FACTOR<br><br>MAX_FACTOR</div><div><br>We can keep default values for them 0.5, as it is in the source code now. When we have MIN_FACTOR == MAX_FACTOR the algorithm will believe as it behaves now. When we have MIN_FACTOR < MAX_FACTOR each time when FACTOR is needed we can select random uniform value in the range. <br><br>What do you think?<br><br></div><div>All the best,</div><div>Todor</div><div><br></div></div></div></div></div></div></div></div></div></div></div></div>