Cloud Service Brokers (CSBs) simplify complex resource allocation decisions, efficiently linking up the tenant dynamic requirements in to providers dynamic offers, where several objectives should ideally be considered. Nowadays, both demands and offers should be considered in dynamic environments, representing particular challenges in cloud computing markets. This work proposes for the first time a pure multiobjective formulation of a broker-oriented Virtual Machine Placement (VMP) problem for dynamic environments, simultaneously optimizing following objective functions: (1) Total Infrastructure CPU (TICPU), (2) Total Infrastructure Memory (TIMEM) and (3) Total Infrastructure Price (TIP) subject to load balancing across providers. To solve the formulated multi-objective problem, a Multi-Objective Evolutionary Algorithm (MOEA) is proposed. When a change arises in the demands or in the offers, a set of non-dominated solutions is found (usually more than one solution), selection strategies were considered in order to automatically select a solution at each reconfiguration. The proposed MOEA and selection strategies, were compared in different scenarios composed by real data from providers in actual markets. Experimental results demonstrate the good quality of the obtained solutions for the proposed scenarios.