This study presents a novel approach named as “multiple objective social group optimization” (MOSGO) to tradeoff time, cost, quality, and carbon dioxide emission (TCQC) factors in generalized construction projects. The proposed algorithm modifies the operation mechanism to balance the exploration and exploitation abilities of the optimization process. The TCQC tradeoff problem considers all types of logical relationships between project activities. Two practical case studies demonstrate the ability of MOSGO-generated, non-dominated solutions. In addition, evidential reasoning is applied to select a compromise solution for project implementation. Comparisons between the MOSGO and four well-known algorithms (MODE, MOABC, MOPSO, and NSGA-II) to verify the efficiency and effectiveness of the developed algorithm. According to the statistical analysis, the proposed algorithm generated the highest values of diversification measurement (DM) of 26.113 and 40.27; the highest values of hyper-volume (HV) of 0.875 and 0.881 in case 1 and case 2, respectively. The proposed algorithm also found solutions with lowest mean ideal distance (MID) and spread (SP) values of 0.872 and 0.462 in the first case and of 0.754 and 0.689 in the second case. MOSGO showed had better diversify and convergence, gained wider spread, and yielded higher uniformity of solutions than the compared multiple objective evolutionary algorithms.

这项研究提出了一种称为“多目标社会群体优化”（MOSGO）的新颖方法，可以权衡广义建设项目中的时间，成本，质量和二氧化碳排放量（TCQC）因素。所提出的算法修改了运行机制，以平衡优化过程的探索和开发能力。TCQC权衡问题考虑了项目活动之间的所有类型的逻辑关系。两个实际案例研究证明了MOSGO生成的非支配解决方案的能力。此外，采用证据推理来为项目实施选择折衷解决方案。将MOSGO与四种著名算法（MODE，MOABC，MOPSO和NSGA-II）进行比较，以验证所开发算法的效率和有效性。根据统计分析，所提出的算法产生的多样性测度（DM）的最大值分别为26.113和40.27；在案例1和案例2中，超体积（HV）的最大值分别为0.875和0.881。所提出的算法还找到了在第一种情况下具有最低平均理想距离（MID）和扩展（SP）值分别为0.872和0.462以及在第二种情况下分别为0.754和0.689的解决方案。与比较的多目标进化算法相比，MOSGO表现出更好的分散性和收敛性，获得了更广泛的传播，并产生了更高的解决方案统一性。所提出的算法还找到了在第一种情况下具有最低平均理想距离（MID）和扩展（SP）值分别为0.872和0.462以及在第二种情况下分别为0.754和0.689的解决方案。与比较的多目标进化算法相比，MOSGO表现出更好的分散性和收敛性，获得了更广泛的传播，并产生了更高的解决方案统一性。所提出的算法还找到了在第一种情况下具有最低平均理想距离（MID）和扩展（SP）值分别为0.872和0.462以及在第二种情况下分别为0.754和0.689的解决方案。与比较的多目标进化算法相比，MOSGO表现出更好的分散性和收敛性，获得了更广泛的传播，并产生了更高的解决方案统一性。