招聘和求职是人才市场上需方和供方之间的博弈过程。鉴于招聘企业与求职人才均无法充分知晓对方的需求信息，招聘网站服务商应运而生。本文引入网站简历适配度参数，构建网站服务商和招聘企业之间的广告合作模型，预测招聘企业、服务商和服务供应链期望效用。招聘企业可选定一个或几个网站服务商构建长期合作的招聘服务供应链，通过合作发布招聘广告的形式，吸引求职人才投递简历，提高招聘效果。研究发现，网站广告效应越大，越便于求职人才投递简历，招聘效用越大；网站简历适配度越大，人才简历量越多，招聘效用越大。预先确定广告赠送率或职位价格，可预测另一参数及其对应的招聘效用。通过案例研究，本文方法成功帮助一家招聘企业选定两个理想的网站服务商构建长期合作的并联招聘服务供应链。实践证明本文广告合作模型具有良好的应用价值。

Recruitment and job hunting is a game process between recruitment enterprises (RE) and job hunting talents (JHT) in talents and human resource market. Since neither of two parties can get full information from the other side, some recruitment website servicers (RWS) appeared. But few studies have addressed that how to select the best suitable RWS.
In this study, a website advertising cooperation model between service providers and recruitment companies is developed which can be used to forecast expected utility of recruitment enterperises, service providers and service supply chain.
Recruiters can select one or several website recruiting service providers and build a service supply chain for long -term cooperation on recruitment (RSSC) which will be helpful for the ability complementary, advertising cooperation,utility maximum and profit sharing amongst these partners. The recruiters and website recruiting service providers can attract talents to submit resumes and improve the effect of recruitment by cooperating and releasing advertisements. In a RSSC, different RWS offers talents resumes from different industries, and resumes matching rates differ from each other according to RE's requirements.
A parameter (ρ_i) is introduced for RE's preliminary judgment on website resumes fit from different RWSs. There are six independent variables:ρ_i(website resumes fit),x_i (advertisement intensity), r_i(advertisement effect), t_i(discounting rate of advertising), w_i(cost of vacant position), and p_i(searching price of vacant position). There are four dependent variables such as y_i(the amount of expected talents resumes),u_Ei(the effect of RE), u_Wi(the effect of RWS), u_Si(the effect of RSSC). Supposing y_i is a decreasing function of p_i, we will have increasing function of r_i and increasing function of x_i with diminishing-marginal effect, i.e. y_i(p_i,x_i)=ρ_i(y_ij-αp_i+r_i√x_i). A Steinberg game is designed in which RWSs make decision first and then REs make decision. With the method, 6 propositions are found for the reference of RE's decision making.
The results of our study show that when the effect of website advertisement is bigger, which makes it more convenient for the talents to submit their resumes, the recruitment effect of RE will be better. When website resumes fit is bigger, which bringing more resumes of talents, the effect of recruitment will be better. If the discounting rate of advertising and the price of positions are redetermined, the other parameters and the corresponding recruitment effect can be predicted.
A case study is provided in which our model was used to assist an enterprise successfully select two ideal website service providers and build a parallel supply chain for long-term recruitment service. All the data are from practical investigation. It proved that our proposed model has practical application value in real world.