Particle Swarm Optimization for Multi-Objective Web Service Location Allocation

Evolutionary Multi-Objective Optimization for Web Service Location Allocation Problem

“…They formulated the problem as a stochastic mixed integer program and proposed a simulation-based hybrid heuristic to solve the dynamic problem under different response time service level. In contrast, in our previous studies [11], [12], we proposed a multi-objective algorithm with linear aggregation using PSO and a multi-objective algorithm with Pareto front using NSGA-II. Both results show that multi-objective model suits the problem well.…”

“…To address discrete problems, Kennedy and Eberhart developed a binary PSO [23]. We applied BPSO in WSLAP in previous study [11] and used the linear aggregation (weighted sum) approach to transform the multiple objective values into a single aggregated value. BPSO can provide a single solution and therefore it is used in the scenarios where a service provider knows their budget and required service response time.…”

“…The threshold plays an important role in searching for solutions for a given problem [11]. The static rounding function has the following drawbacks.…”

“…To evaluate the performance of our proposed BMOPSOCD with dynamic function, we conducted experiments to compare its performance with BNSPSO, BPSO from our previous research [11], and NSGA-II from [12]. Based on the analysis in Section 5.5, we use the Quadratic function as the rounding function.…”

“…To address WSLAP as a multi-objective problem, we first proposed an aggregation approach with Binary Particle Swarm Optimization (BPSO) [11]. Even though the BPSO approach can solve the problem, a single-objective algorithm can only provide one solution for each run, hence, BPSO cannot provide alternatives when service providers do not have preferences.…”