Privacy-Enhanced Multitasking Particle Swarm Optimization Based on Homomorphic Encryption

evolutionary multitasking optimization (EMTO) is a new optimization paradigm proposed in the field of evolutionary computation in recent years. EMTO can solve several different optimization tasks simultaneously and facilitate superior convergence characteristics by transferring effective knowledge among the tasks. However, existing EMTO usually focuses only on facilitating convergence characteristics while neglecting the potential privacy leakage problem in the knowledge transfer between different tasks. The privacy leakage could result in considerable financial losses or severe reputation impairment, which may impede the development of EMTO in real-world applications. To solve the problem of privacy leakage in EMTO, this article proposes a privacy-enhanced multitasking particle swarm optimization algorithm. A knowledge transfer strategy with privacy preservation is designed based on homomorphic encryption by combining multitasking particle swarm optimization. In addition, an intertask knowledge transfer mechanism implemented in a low-dimensional subspace is introduced to reduce the extra computational burden caused by privacy preservation. Comprehensive experiments are conducted on synthetic and NAS problems to verify the effectiveness of the proposed method. According to the experimental results, the proposed method has remarkable advantages in privacy preservation compared to existing EMTO.

I. Introduction

Evolutionary algorithm (EA) have been widely used as general optimizers for various problems due to their strong problem-solving abilities [1]. The design of traditional evolutionary algorithms EAs usually focuses on efficiently solving a single optimization problem. However, since problems with relevance are prevalent in the real-world applications, EAs should be equipped with the ability to accelerate search by leveraging the complementary experience of related tasks. Consequently, Gupta et al. first proposed the concept of evolutionary multitasking optimization (EMTO) [2], [3]. Different from traditional optimization search, EMTO is able to utilize intertask knowledge transfer to perform the evolutionary search for different optimization problems simultaneously. The knowledge transfer in EMTO represents the crossover of genetic material or the sharing of search experience between different tasks, which can be performed implicitly or explicitly. The effectiveness of EMTO has been preliminarily verified by proposing a multifactorial EA (MFEA) in [3]. In the multitasking environment, MFEA uses a single population to search for the optimal solutions for each task. Inspired by the biocultural models of multifactorial inheritance, cultural effects in MFEA are incorporated via two features of multifactorial inheritance acting in concert, namely, assortative mating and vertical cultural transmission.