A Clonal Selection Optimization System for Multiparty Secure Computing

Algorithm 4

Federated learning process.

	Initial:
	Parameters function , d is the maximum engagement of each client, the maximum iterations of node is Imax
	Federated central sever do:
	For each weight of model do
	For each iteration do
	Local optimization process:
	Select a fixed number of eligible clients
	For each clients do
	Optimize parameters with local data:
	For client Ck: Client parameter update (Ck, θi) do
	Divide the dataset DA into several subdatasets Ds
	For local iteration do
	Update
	Return θ to the federated central server